2008 Annual Report Conference on Electrical Insulation Dielectric Phenomena

Inclined-Plane Tracking and Erosion Test

according to the lEG 60587 Standard
A. Krivda, L.E. Schmidt, X. Kornmann, H. Hillborg', H. Ghorbani, A. Ghorbandaeipour and M. Eriksson
ABB Switzerland Ltd, Corporate Research
Baden-Daettwil, Switzerland
'ABB Sweden AB, Corporate Research
Vasteras, Sweden
Figs. 3 and 4 show new and used tracking and erosion
electrodes. It can be seen that after a six hours long test at 4.5
kV the electrode is severely ablated. To ensure statistically
relevant results without any memory effect from the past tests
ABB uses new electrodes for every test.
Figs. 5 and 6 show detail of the damaged electrode shown in
Fig. 4 before and after cleaning procedure with sand paper.
This is a method used by some material suppliers to recycle
the test electrodes. Decomposed material debris and
roughened surface can be seen in Fig. 5 while after cleaning
the surface became much smoother. The cleaned surface is,
however, of different quality than the original one shown in
Fig. 2. This may affect future tests because the smooth surface
will induce different electric fields and electron emission rates
than the rough surface.
To avoid the electrode effect an electrode-less set-up should
be designed and tested. In this set-up polymeric surfaces with
sufficiently high surface fields rather than metallic electrodes

Abstract- The main goal of the paper is to summarise

experiments using the tracking and erosion test according to the
IEC 60587 standard. Various polymeric materials, e.g., high
temperature vulcanised silicone rubber, liquid silicone rubber,
epoxy resin and fibre reinforced materials were tested and
differences in performance noted. A leakage current monitoring
system has also been developed and the results of leakage current
analysis are presented. Possible improvements of the present
standard are also discussed.
I. INTRODUCTION

Accelerated testing of outdoor polymeric materials with a

reliable future performance indication is still an elusive task.
Although several standards for materials testing are available
[1-7], at best they provide a comparative performance index
rather than the absolute future performance in actual power
equipment which is also affected by electrical design and not
only by material selection. The main goal of the paper is to
summarize more than two hundreds experiments carried out at
the ABB Corporate Research Center in the past two years and
to initiate discussion on some critical issues concerning the
IEC 60587 standard.
II. EXPERIMENTAL PART

The experiments in this study were performed using a

commercially available Furukawa tracker. The air in the test
chamber was continuously extracted to prevent condensation
of water vapor on samples. New electrodes were used in every

test.

III. ELECTRODES

Neither the manufacturing method nor the number of tests

which can be performed with the same electrode are specified
by the present standard [1]. Figs. 1 and 2 show scanning
electron microscopy (SEM) photos of new electrodes provided
by two different suppliers. The surface quality is very different
(smooth vs. rough) and it will affect the electric field around
the electrodes as well as the emission of electrons from the
electrodes. Comparison of results between ABB and various
suppliers has shown that the rough electrode surface is able to
differentiate between materials far better than the smooth
electrode surface. Future round-robin tests are planned to
quantify the effect more precisely.

978-1-4244-2549-5/$25.00 ) 2008 IEEE

rig.

263

D.

NCWeweectroUe 10or traCKing dani CIerosiol test.

Fig. 9. High temperature vulcanized silicone rubber tested at 4.5 kV - set Ic.

V. TEST VOLTAGE LEVEL

The present standard specifies four different test voltage

levels - 2.5 kV, 3.5 kV, 4.5 kV and 6 kV. In the experiments
performed in this study, the 4.5 kV voltage level appeared to
be the most critical as arcing tend to concentrate only on one
or two points where it caused severe local damage in the form
of extensive erosion. In contrast to the 4.5 kV experiments, in
the 6 kV tests the arc tended to move all over the sample
surface. It caused only a limited local damage and materials
thus passed the test at the 6kV voltage level.
Some materials exhibited a clear performance step when
tested at different voltage levels. Figs. 10 to 13 show two
different silicone materials tested at 3.5 kV and 4.5 kV. It can
be observed that both materials performed satisfactorily at 3.5
kV but one of the materials failed completely the test at the 4.5
kV voltage level.
In a recent paper [8], the authors have studied the erosion
depth and mass loss at several voltage levels concluding that
the 4 kV voltage level caused the most damage. While it is of
definite interest to perform such studies the results have to be
viewed with caution. To ensure full statistical independence of
results, new electrodes and at least fifteen samples should be
used in such studies.

D.Detail of electrode after six hours test at 4

after cleaning procedure with sand paper.

IV. NUMBER OF TEST SAMPLES

To ensure statistically relevant results the present standard

specifies to test five samples. As our experiments proved this
criterion may not be valid for all types of materials. Figs. 7, 8
and 9 show the results of three tracking and erosion tests. The
samples were prepared from the same material batch and new
electrodes manufactured by the same supplier were used in the
tests. The inconsistent performance of the material can clearly
be seen. The ABB internal standard for material acceptance is
to test at least fifteen samples to ensure future reliable
performance of the material.

264

perform reliably

on some heavily polluted locations where

silicone is a clear winner.
Whatever the future erosion criterion will be, nonmechanical
means of measuring erosion depth, for example stereoscopic
imaging, should be used in future as the standard depth gauge
may compress soft material samples and distort the
ittcIhmeasurement by several tenths of a m-iillimeter.

VII. LEAKAGE CURRENT MONITOR

To carry out detailed analysis of the tracking and erosion test,

leakage current monitoring system was developed. The
sampling frequency was about 10 kHz, the total sampled time
can be varied between two to five seconds and two to three
measurements per minute were performed. The following
parameters were derived from the leakage current: the
absolute peak per sample period, the accumulated charge per
sampled period, harmonic and current distributions as a
function of the total test time.
Several observations can be made. The leakage current
increased before the electrical breakdown, see Fig. 14.
Although there were many individual peaks - the trends rather
than individual peaks are important.

Fig. 10. Silicone material A tested at 3.5 kV.

Fig. I11. Silicone material B tested at 3.5 kV.

Fig. 12. Silicone material A tested at 4.5 kV.

Fig. 14. Absolute peak in mA of leakage current measured during three

second period and shown as a function of the total test time.

The accumulated charge per sampled period, see Figs. 15 and

was a useful indicator of the start of a deep erosion
process. A clear step in the accumulated charge value can be
observed at the beginning of deep erosion process.

16,

Fig. 13. Silicone material B tested at 4.5 kV.

VI. EROSION DEPTH

The present standard does not specify the maximum

permitted erosion depth. The only criterion is that the samples
must not be perforated at the end of the test and the erosion
depth should be noted for every sample. Some utilities,
however, require that the erosion depth is not greater than a
specific value, for example. 2.5 mm. It is difficult to assess the

IA

merits of this criterion - it may work well for some materials

and fail the other. For example, porcelain may not be eroded
during the test yet as a material it may completely fail to

t1.I.efy--

Fig. 15. Accumulated charge per sampled time measured during a

three seconds period and shown as a function of the total test time.

265

r1i. 1U.

111Ul1l IlUDUlI Saldll

l
lS
tiLO
LU-1LSjJUIIIgU

\aKag

LLt11LS SlIUWlI

in Fig. 15 (1 = failed electrically, 3 = passed, 5 = passed electrically, failed by

the erosion depth criterion).

Visual comparison of fast Fourier transform data obtained

from the leakage current showed a higher content of higher
harmonics in the more damaged samples, see Figs. 17 and 18.
Detailed mathematical analysis is required to study subtle
differences between the failed samples and samples which

,ig. iv. reaK ieaKage clurret CtISrLDUtlonas a ilUlloULn Ui tUtai test timc. 111n
start of the deep erosion process is indicated with an arrow and results in
perforation of the sample similar to the sample #1 shown in Fig. 16.

17. Harmonics ot leakage current tor the worst and best sample.

The current histogram as a function of time also proved to be

a useful indicator of the start of deep erosion process, see Fig.
19. At the start point of the deep erosion process the low level
currents ("noise") have almost disappeared and high level
currents started to dominate.

(the photo is rotated by 90 degrees).

In fiber-reinforced materials the damage propagated along

the path of the least resistance - the fiber/polymer matrix
interface, see Fig. 20. The continuous flow of the conductive
solution caused sometimes inversion of the electric field
which resulted in the backward growth of tracking in
randomly oriented fiber samples, see Fig. 20.

VIII. MATERIAL TESTING

During the course of experiments more than one hundred

different materials were tested: silicone rubber, epoxy, fiberreinforced and thermoplastic materials.

266

When some thermoplastic materials were tested, see Fig. 21,

the main cause of damage was actually decomposition by heat
from resistive current heating the conductive solution and not
from dry band arcing.
The type of filler has a significant influence on the type of
erosion path, see Figs. 22 to 24 which show various silicone
rubber materials after the tracking and erosion test at 4.5 kV.
IX. CONCLUSIONS

The following issues should be discussed in preparation of

the future tracking and erosion standard:
1) The production methods for electrodes (laser cutting, water
jet cutting).
2) The use of new electrodes in every test.

3) The acceptance criteria - fifteen instead of five samples.

4) The erosion depth measurement by optical rather than
mechanical methods.
REFERENCES
[1]

Electrical Insulating Materials Used under Severe Ambient Conditions Test Methods for Evaluating Resistance to Tracking and Erosion, IEC
60587 International Standard, 3rd ed., 2007.

[2] Method for the Determination of the Proof and the Comparative
Tracking Indices of Solid Insulating Materials, IEC 60112 International
Standard, 4th ed., 2003.

[3]

Standard Test Method for High-Voltage, Low-Current, Dry Arc

Resistance of Solid Electrical Insulation, ASTM Standard D495-99,
1999.

[4]

Electrical Insulating Materials - Method to Evaluate the Resistance to

Tracking and Erosion - Rotating wheel Dip Test, IEC 1302 International
Standard, 1st ed., 1995.

[5]

Composite Insulators for AC Overhead Lines with a Nominal Voltage

Greater Than ]OOOV - Definitions, Test Methods and Acceptance
Criteria, IEC 61109 International Standard, Ist ed., 1992.

[6]

Surge Arresters - Metal-Oxide Surge Arresters without Gaps for AC

Systems, IEC 60099-4 International Standard.

[7]

Composite Insulators - Hollow Insulators for Use in Outdoor and

Indoor Electrical Equipment - Definitions, Test Methods, Acceptance
Criteria and Design Recommendations, IEC 61462 International
Standard, 1st ed., 1999.

[8] J.M. Seifert, R. Barsch and W.L. Vosloo, "Dimensioning of the housing
profile of silicone rubber composite insulators for harsh marine pollution
conditions", CIGRE, Paris, France, Paper D1-303, 2008.

267