OE1-03

ELECTRICAL CHARACTERISTICS AND PROPERTIES OF A STUB

(DAMAGED TOUGHENED GLASS INSULATOR)
1* 1 1 2
JM George , S. Prat , S. Tartier and Z. Lodi
1
Sediver S.A. R&D St Yorre, France
2
Sediver Canada, Montreal, Canada
*Email: <[email protected]>

Abstract: Overhead line insulators can be damaged for various reasons during their
service life. Porcelain or composite insulators once damaged can display a large variety
of conditions and aspects. On the other hand, damaged toughened glass insulators
always appear as a stub thanks to the properties of toughening. While the mechanical
residual strength of a stub is covered by standard tests, its electrical performance is not
always fully understood by the users. This paper presents a comprehensive approach to
the electrical characteristics of a glass stub taken individually or in a string of insulators. A
model of a stub is being proposed in an equivalent electrical circuit. Test results on stubs
tested individually or in full strings are analysed and will help to understand their
behaviour in situations in line with normal service conditions as well as during over
voltages. The influence of environmental elements (dry, wet or humid conditions) is also
presented through selected pre-stress conditions. Finally, the electrical behaviour and
characteristics of a stub are studied after mechanical and thermal mechanical
preconditioning. The electrical properties of a stub offer a significant contribution to the
utilities in their maintenance, inspection and live line work strategy.

1 INTRODUCTION instantly released when any overwhelming damage

is done to the dielectric.
While the performance of overhead line insulators
is well described through standard test procedures,
there is much less information relative to the
performance of damaged insulators. In reality the
only standard test existing in this category is the
mechanical residual strength test which exists with
various procedures for cap and pin suspension or
tension insulators [1] [2]. Unfortunately, given the
large diversity of possible damage modes on
polymers such an approach does not exist for
composite insulators.

The results described in this paper do not focus on Figure 1: Typical aspect of a toughened glass stub
mechanical aspects of damaged toughened glass
insulators, already largely described [3], [4], but As a result, and unlike porcelain which can break
more specifically on the electrical performance and in large pieces with randomly distributed cracks
behaviour of a broken glass insulator. (figure 2a), a broken glass insulator systematically
becomes a stub (figure 2b). The aspect and
The electrical characteristics and main parameters properties of stubs can therefore be predicted with
of a stub (broken glass shell) are defined. Electrical consistency given the reproducibility of the end
properties of a stub have been evaluated as a condition of the insulator once damaged [5], [6].
single unit as well as part of a complete string
through tests simulating the diversity of stresses 3 MAIN ELECTRICAL CHARACTERISCS OF
which can be encountered in service. A TOUGHENED GLASS INSULATOR

2 BRIEF DESCRIPTION OF A STUB The characteristics of a cap and pin insulator

dielectric are mainly defined through the electrical
The common designation of “stub” (figure 1) resistance (R) and capacitance (C). Unlike for
applies to any toughened glass insulator damaged porcelain, the nature of glass (amorphous material)
on the shell. The thermal toughening of the glass produces inherently very stable values since it is
shell during manufacturing introduces a not affected by micro cracks, micro structures and
mechanical pre stress in the glass which will be crystallographic structure or defects which cannot
exist in toughened glass dielectric [7].

1154
 OE1-03

(kN)

a: Porcelain discs b: Toughened glass

Figure 4: Example of capacitance values of
Figure 2: Damaged cap and pin units in a string. toughened glass insulators.

In this regard, and unlike for porcelain dielectrics The surface resistance of the insulator (meaning
the values of R and C are strictly defined by the the electric resistance along the surface of the
chemistry of the glass itself and the size of the skirt) is largely a function of the environmental
dielectric (geometric considerations like thickness conditions; this surface resistance is considered as
of the head in a cap and pin insulator) and are not variable.
affected by ageing or time.
Concentrating strictly on dielectric material
Figure 3 gives an example of the body resistance behaviour we will not focus on this variable
of similar toughened glass insulators for AC resistance. The equivalent circuit in dry and clean
applications and DC applications. To be correctly conditions can therefore be represented by 5a, and
measurable, the resistance is tested at high to the extent that the body resistance is very high,
temperature. The resistance of a typical insulator cap and pin insulator units can be simplified as
unit at ambient temperature can be estimated from represented by 5b.
10GΩ to 500GΩ for AC and DC glass respectively.
The difference in values between AC and DC is
strictly the result of the bi-alkali effect in the
chemistry used for the DC glass.
C0 R0 V C0 V

a) b)

Figure 5: Equivalent circuit of a cap and pin

suspension insulator. a) Full circuit. b) Simplified
circuit.

4 STUB SCHEMATIC ELECTRICAL CIRCUIT

Doing the same exercise with a stub requires the

understanding that, once broken, the remaining
glass stub will not see internal arcing during a
voltage increase but rather flashover by an arc
Figure 3: Example of body resistance measured at externally crossing the section of broken glass
high temperature for AC and DC insulators as per between the cap and the pin as can be seen in
IEC 61325 [8] [9]. figure 6. This external air gap is short compared to
the longer distance for an arc to find its way
The value of the capacitance of a toughened glass internally between the glass fragments
insulator is almost constant and is only slightly compressed between the cap and the pin once the
dependent on the size of the glass head inside the insulator is broken.
cap.

A median value of C around 70pF can be

considered for a complete glass insulator
corresponding to a mechanically mid-range
product, as can be seen in figure 4.

1155
 OE1-03

The following section describes the values of C

and R of a stub under various preconditioning
circumstances as well as the stub behaviour.

5 INFLUENCE OF ENVIRONMENTAL
STRESS CONDITIONS TO THE
ELECTRICAL PROPERTIES OF A STUB

The electrical behaviour of a stub was determined

Figure 6: Flashover across the remaining section after various pre-stress conditions as explained
of broken glass of a stub. (Left AC, right DC). below. For each case the stubs were tested after
being subjected to one or the other preconditioning
The thickness of the glass is slightly variable but in parameters (figure 8). The tests presented in the
a narrow range (we will consider 12mm as a following sections will cover most frequently asked
reference value), we can assimilate this external questions relatively to the electrical behaviour of
and constant distance to an air gap operating stubs in service.
precisely as such. Tests have shown that 12kV is a
typical discharge value for this phenomenon in AC
in dry and clean conditions. For DC average values
around 14kV were found for both polarities. The
capacitance and resistance of the stub will be
changed by the fact that the glass is broken. We
can therefore consider that both C and R will
change leading to the equivalent electric circuit
Figure 8: Possible stresses for a stub in service.
shown in figure 7.
5.1 Atmospheric conditions
Cs is the additional capacitance and Rs the
resistance arising from the interfaces in the broken
glass inside the head of the insulator. The 5.1.1 Stubs in water
equivalent capacitance of a stub is thus defined by:
Stubs were immersed in salt water for one hour.
C = Co + Cs and 1/R = 1/Ro +1/Rs Electric tests were performed 2 hours after being
removed. All stubs flashed externally.
Where C, Co and Cs = capacitance in Farad (F)
and R, Ro and Rs= resistance in Ohm (Ω) Additionally, the electrical resistance of the stubs
was measured highlighting a reduction of the
resistance to around 0.1 MΩ compared to the initial
C = 200 pF; R = 20 MΩ
value of 20 MΩ while the capacitance increased to
reach values around 600 pF compared to the initial
These values are defined for normal atmospheric
value of 200 pF.
conditions in clean ambient laboratory conditions. It
is extremely important to understand the behaviour
5.1.2 Stubs in dry air conditions
of the air gap corresponding to the distance over
the glass between the cap and the pin at the
bottom of the stub. This distance is constant and Samples of stubs were preconditioned for 24h in a
represents the shortest distance for an arc to take climatic chamber at 40°C and 22% humidity. All
place. While for a complete insulator the flashover samples were electrically tested confirming an
is defined by the arcing distance, in the case of a external flashover. Figure 9 shows the evolution of
stub the arcing value is strictly defined by the air C and R at various times after being introduced
gap between the internal edge of the cap and the from ambient conditions into the climatic chamber.
cement around the pin. It is interesting to note that during the drying
process the resistance is slightly increased while
the capacitance is slightly decreased. Despite
these changes it is important to note that the
behaviour of the stub is not changed, always
showing systematic external arcing.
C0 CS R0 RS 12 kV V
5.1.3 Stubs under icy conditions

Stubs have been frozen for 24h after being

immersed in water for 48 h The stubs were
Figure 7: Equivalent electric circuit of a stub in dry electrically tested immediately after being removed
and clean conditions from the freezer, and all stubs flashed externally

1156
 OE1-03

after the leakage current on the surface of the stub In a final stage, all stubs were electrically tested
had melted the ice. under a permanent mechanical load of 70%, again
leading to systematic external flashover.

5.3 Electric stresses on stubs

5.3.1 Steep front tests on stubs

Stubs were subjected to a steep front wave test

which was as severe as possible given the short
air gap on the stub (figure 11). A steepness in the
range of 1000 kV/µs was reached during these
tests. All stubs flashed externally. Additionally after
Figure 9 : Evolution of R and C as a function of the steep front test the stubs were subjected to a
time after introduction of a stub into a dry air 3h withstand test at 11kV as shown in figure 11. All
climatic chamber. the stubs in this test displayed systematic external
flashover in a retest after this sequence.
5.2 Mechanical pre stresses on stubs

While the mechanical strength of stubs has been

largely demonstrated over the last decades and
clearly described through standard tests [1], [2],
some additional mechanical stresses were
considered.
a) b)
5.2.1 Vibration stresses
Figure 11: (a) Steep front on stubs and (b) power
frequency withstand test after steep front test.
Stubs were tested after having been subjected to
11 million cycles where the applied load varied
5.3.2 Power arc on stubs
from 40% to 50% of the mechanical rating of the
insulators at a frequency of 12Hz. All the stubs
Stubs were subjected to a power arc following the
were tested electrically and in all cases the arc was
set up and test conditions of IEC 61467 [11]. The
observed in the air gap outside the stub. (A
tests were performed simulating various
verification of their mechanical integrity showed
configurations of a string made of 6 units. The stub
that all failing loads were above mechanical rating).
was placed once at the bottom, once in the middle
and once at the top of the strings. Each string was
5.2.2 Thermal mechanical stresses
submitted to two consecutive shots of 6kA/0.2s.
The same test was performed with strings which
Full strings of stubs have been tested in thermal stubs had previously been immersed for 48h in a
mechanical conditions corresponding to the most 10g/l salt solution. An additional test of 6
severe testing conditions [10] as shown in figure consecutive shots was made on a string
10. constituted only of 6 stubs.

The choice of the applied current was made to

ensure that the test would optimize the chances to
keep the arc on the surface of the stub rather than
moving away from the insulator string. This detail
reinforces the severity of the applied stress on the
stub itself.

Figure 12 shows the tested strings. Figure 13

shows the string made of 6 stubs during the power
arc. The presence of 6 separate arcs crossing the
external distance outside the stubs indicates
Figure 10: Left: Thermal mechanical test on stubs. clearly that the stubs do not behave like punctured
Right: electromechanical test at 70% of rating. discs, and as such illustrates clearly the difference
with porcelain discs which, once punctured, could
Applied load was 70% of the rating, temperature lead to a string separation resulting from internal
cycles were from -50°C to +50°C. After thermal arcing along the puncture path leading to
mechanical stresses were applied all the stubs excessive energy accumulation inside the cap.
were electrically tested resulting in systematic
external flashover.

1157
 OE1-03

complete insulators falls below what is required to

withstand the operating voltage. However, with half
of the string broken the operation of the line (80kV
phase-ground) is still maintained (except for cases
where pollution is present, for which leakage
distance is the prevailing factor). These tests were
performed with line end units 1 and 2 always
broken.

Figure 12: Strings with stubs after power arc

testing.

Figure 14: Withstand values of a 138kV string with

stubs.

6.2 Switching impulse on 500kV strings

Figure 13: String of stubs during a power arc test At 500kV the switching impulse behaviour
showing systematic external arcing. becomes a major focus for line performance.

6 ELECTRICAL PROPERTIES OF STRINGS

CONTAINING STUBS

The predictability and strength of a damaged

toughened glass insulator has been largely
demonstrated over the past decades from a
mechanical point of view [3].

From an electrical point of view as well, there is no

hidden risk or unpredictable behaviour to fear [12],
making toughened glass insulators the preferred
type of insulator from a live line work point of view Figure 15: U50 dry switching impulse performance
[12]. Specific rules relatively to clearance distance of 500kV string with 5 stubs (A: full string B: 5
and national standards differ from one country to stubs in 3-7-11-15-19 from line end C: 5 stubs in 1-
another in terms of number of damaged units 7-11-15-19 D: 5 stubs in 1-7-11-15-22).
allowed in a string [13] before a maintenance
action is required. The tests were performed with a string of 22 units
ref. B12/140. A string having 5 stubs distributed in
The following aspects of performance of strings positions 1-7-11-15-22 was tested under dry
with stubs have been selected to form a rational switching impulses. Figure 15 shows the
understanding of the string behaviours. comparative results between this configuration and
a complete intact string. It is interesting to note that
6.1 Dielectric tests on 138 kV strings the performance of a string containing about 22%
of stubs maintains about 90% of its original
Typical 138 kV strings of 8 units (ANSI class performance.
52-5) were tested under lightning impulses as well
as dry and wet power frequency. Intact units were 6.3 RIV on 500kV strings
replaced with stubs in various locations along the
string. The results in figure 14 show that the most The following RIV test was performed according to
critical case is the wet power frequency ANSI C29-1 [14] on a 500kV string made of 25
performance. When 6 out of 8 units are in a stub ANSI 52-5 glass units, equipped with normal
condition (75 % of the string being broken), the hardware configuration, having a stub at the live
remaining leakage distance of the two last

1158
 OE1-03

end. At nominal voltage the RIV levels are similar while ensuring the most reliable service conditions.
both with or without a stub. The absence of urgency resulting from the
behaviour of stubs makes toughened glass the
most effective solution for live line work and full
cycle cost.

8 REFERENCES

[1] IEC 60797

[2] ANSI C29.2
[3] A. Mishra, R. Gorur, S. Venkataraman:
“Evaluation of Porcelain and Toughened
Glass Suspension Insulators Removed
from Service”, IEEE Transactions on
Dielectrics and Electrical Insulation,
Figure 16: RIV results of a 500kV string containing Vol. 15, N°2, pp.467-475, 2008
a stub in the glass string at the live end. [4] JM. George, E. Del Bello: “Assessment of
electrical and mechanical performance of
The actual behaviour of a string containing a stub toughened glass insulators removed from
is the result of the voltage distribution along the existing HV lines”, CIGRE Meeting Calgary,
string. The voltage distribution measured on a Paper N° 451A, 2007
500kV string composed of 20 glass ANSI 52-5
class insulators is shown in Table 1. [5] R. Gorur: “A review of Insulator Failures”,
INMR, Vol.19, N°4, pp.66-77, 2011
[6] IEEE Lightning and Insulator
Subcommittee WG on Insulators TF on
Insulator Loading: “High Voltage Insulators
Mechanical Load Limits-Part I: Overhead
Line Load and strength Requirements”,
IEEE Transactions on Power Delivery,
Vol.27, N°3, pp.1106-1115, 2012
[7] A. Rawat, R. Gorur: “Microstructure Based
Evaluation of Field Aged and New
Table 1: Percent voltage distribution of a 500kV Porcelain Suspension Insulators”, IEEE
string of toughened glass insulators with a stub in Transactions on Dielectrics and Electrical
various positions. Insulation, Vol.16, N°1, pp.107-115, 2009
[8] IEC 61325
It is interesting to note that during this test there
was no partial arcing on the stub and the voltage [9] JM. George, Z. Lodi: “Design and selection
distribution is similar to a string with only intact criteria for HVDC overhead transmission
discs everywhere else. The drop of potential of the line insulators”, CIGRE Conference on
stub in various positions is stable for whatever Power Systems Toronto, Paper N°189,
position in the string. This is the result of the 2009
difference of the capacitance of a stub versus a [10] CSA C411.1
complete disc (about 4 times more for a stub as
explained earlier). [11] IEC 61467
[12] R. Gorur, D. Shaffner, W. Clark, R. Vinson,
7 CONCLUSION D. Ruff: “Utilities share their insulator field
experience – Four utilities discuss the
The geometry and nature of stubs provide unique advantages and disadvantages of
properties to a damaged toughened glass porcelain, toughened glass and polymer
insulator. The benefit of repeatability in the insulators”, Transmission & Distribution
condition of a stub and the specific values of World, N°4, pp.17-25, 2005
resistance and capacitance make that over
voltages under the variety of circumstances [13] ESMOL: “Minimum Number of Good
encountered in service will systematically generate (Healthy) Porcelain or Glass Insulator
an external arc and not a puncture as may be the Units in a string for Live Work”, IEEE
case with porcelain. The predictable and safe Transactions on Power Delivery, Vol.17,
behaviour of a stub in a string allows maintenance N°3, pp.809-814, 2002
crews to forecast easy and low cost maintenance [14] ANSI C29.1

1159