See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/224131365

A New Approach to Anti-fog Design for Polymeric Insulators

Article  in  IEEE Transactions on Dielectrics and Electrical Insulation · May 2010

DOI: 10.1109/TDEI.2010.5448087 · Source: IEEE Xplore

CITATIONS READS
13 211

7 authors, including:

A. Haddad Ronald Thomas Waters

Cardiff University Cardiff University
224 PUBLICATIONS   1,236 CITATIONS    106 PUBLICATIONS   953 CITATIONS   

SEE PROFILE SEE PROFILE

H. Griffiths Krystian Leonard Chrzan

Khalifa University Wroclaw University of Science and Technology
157 PUBLICATIONS   773 CITATIONS    78 PUBLICATIONS   371 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

uprating of overhead lines View project

Transformer diagnosis using Frequency Response Analysis View project

All content following this page was uploaded by Krystian Leonard Chrzan on 20 May 2014.

The user has requested enhancement of the downloaded file.

IEEE Transactions on Dielectrics and Electrical Insulation Vol. 17, No. 2; April 2010 343

A New Approach to Anti-fog Design

for Polymeric Insulators
A. Haddad, R. T. Waters, H. Griffiths, K. Chrzan, N. Harid, P. Sarkar
and P. Charalampidis

High Voltage Energy Systems Group,

School of Engineering, Cardiff University,
The Parade, Cardiff CF24 3AA, UK

ABSTRACT
Electric field strength and leakage current density in the shank region of a polluted
polymeric insulator may cause local dry-banding, with a risk of long-term degradation
from partial-arc discharges. In a novel approach to dry-band control, the
characteristics of silicone rubber surfaces with a textured finish are investigated and,
depending on the geometry chosen for the texture, two useful objectives can be
achieved. First, the increase of surface area can both reduce the leakage current density
in the vulnerable shank region, and also increase the longitudinal creepage distance.
Secondly, the damage arising from surface discharges can be mitigated. The principles
of the proposed anti-fog design are described in detail. While complete prototype
insulators are now being specified for fabrication, this paper reports preliminary
results from tests using rectangular samples, which show an improvement in
performance. The tests are of two kinds: a) inclined-plane, to assess erosion and
tracking on rectangular samples of insulation materials. Such samples have been
prepared both with a conventional plane surface and with a textured surface for
comparative tests; b) clean-fog tests that use an initially dry pollution layer of known
salinity. Here, plane and textured samples have been used for comparative tests, where
leakage current and discharge activity are monitored during the generation of the fog.
Index Terms — Polymeric insulators, silicone rubber insulation, ageing indicators,
pollution, insulator testing, dry-banding, accelerated ageing tests, HV tests.

Polymeric insulators are gaining acceptance for use with

1 INTRODUCTION
THE hydrophobic surface of polymeric insulators, and
particularly the property of many polymeric materials that
enable the transfer of hydrophobicity to an overlying pollution
layer, offers a significant advantage compared with glass or
ceramic units used in overhead transmission and distribution
systems in coastal or industrial environments. The reduction of
wetting by hydrophobicity can reduce leakage current and
discharges that may lead to flashover. Nevertheless, under
severe ambient conditions, dry bands can still form, especially
before hydrophobic recovery can occur in the pollution layer,
and these are known to result in partial-arc discharges in the
shank region of polymeric insulators. The local increase of
electric field strength and leakage current density, in a
pollution layer covering the shank region of the fibreglass
core, is conducive to dry-banding.
Manuscript received on 20 May 2009, in final form 20 September 2009.
overhead lines at distribution voltages up to 132 kV, but concerns
remain about their long-term performance in polluted atmospheric
conditions. A critical region is the recessed area below the shed of
the insulator, where washing is least effective and deposits may
accumulate. The geometry of the insulator shows that the density
of surface discharge currents flowing from one end of the insulator
to the other is highest at the shank of the insulator. As a
consequence, the surface electric field strength is also at a
maximum in this region, and there is a heightened risk of surface
heating, which may cause dry bands, partial arcs on the surface and
consequential thermal damage. This action is limited in glass and
ceramic cap and pin insulators by ‘anti-fog’ designs having
additional rims on the underside of the shed which lengthen the
creepage distance and reduce the magnitude of the leakage current
[1], but because of the nature of the moulding process, this is not
possible in multi-shed polymeric insulators. However, the
moulding properties of polymeric materials have led to a patented
design programme where the pollution properties of a silicone
rubber surface are improved by a textured finish [2].

1070-9878/10/$25.00 © 2010 IEEE

344 A. Haddad et al.: A New Approach to Anti-fog Design for Polymeric Insulators
The texture consists of a regular raised geometrical pattern of a
kind that does not impede the removal of the insulator from the
mould after fabrication. It is shown that, depending on the
geometry chosen for the texture, the advantages are: a) an
increase of surface area that can reduce the leakage current
density in the vulnerable shank region, and b) the textured surface
that results in a substantial increase in the longitudinal creepage
distance, so reducing the stress in specific areas of the insulator.
In this paper, we describe the basic principles of the anti dry
band design for polymeric insulator surfaces that allows better
performance under polluted conditions. A follow-up research
and development programme, investigating the relative
performance of complete plain-surface and textured insulators,
is in progress. The present paper reports on initial tests where
the pollution properties of rectangular samples of a silicone
rubber surface with a textured finish [3,4] are investigated.
Experimental results are presented for silicone rubber samples
with a textured surface consisting of nearly contiguous 2mm
diameter protuberances in a hexagonal array. The preliminary
results for both kinds of tests showed an improvement in
performance for the controlled-texture samples.
These tests with rectangular samples are being continued in
order to optimize the choice of texture for complete insulators.
Subsequent tests with such insulators will assess engineering
design factors such as form factor and specific creepage
length.
2 DRY BAND FORMATION
Under adverse operation conditions for outdoor insulation,
even for hydrophobic materials, an overlying pollution layer
on the polymeric surface can result in a significant leakage
current in light rain or fog. In this case, the structure carries a
longitudinal surface current, I, of which the current density, J,
is non-uniform. This occurs even where the pollution layer is
of uniform volume conductivity, V, and thickness, T, because
the radius R, therefore, the circumference, S, of the circular
surface contours varies along the sheds and shank of the
structure.
The surface electric field strength, E, is also longitudinal
and will be non-uniform. Since the surface regions of greatest
J and E will coincide if the layer conductance VT is uniform,
the heating of the moist pollution layer is significantly greater
in these regions, risking the formation of dry bands. The
power density dissipation P of the surface layer heating is
given by:
P = EJ = J2/V = I2/(VS2T2) (1)
The greatest heating of a uniform pollution layer will occur
on the insulating structure in the region of the smallest contour
perimeter S (min), so that dry bands will most easily form at
the shank of the structure. It has been found that even
polymeric structures frequently fail because of damage in the
shank region, where partial-arc activity is greatest.
The purpose of a textured surface is, therefore, the
reduction of both electric field strength and current density in
vulnerable regions of an insulator, with the aims of both
mitigating dry band effects and the reduction of surface
damage from partial-arc discharges.
The pollution distribution, in terms of both the salt deposit
density (SDD) and non-soluble deposits (NSDD), can be
expected to be more non-uniform for a textured surface. The
wetting of the pollution layer would also be more non-
uniform. However, local areas of high SDD or moisture
content will have increased conductance VT and thus higher
current density J. As a result, the local evaporation rate will be
higher and will reduce the conductance in such regions.
3 TEXTURED SURFACES
3.1 SURFACE PATTERNS
As noted in our patent [2], the surface heating (and thus
drying rate) of any region of the pollution layer that results
from a conduction current of local density J and electric field
strength E will be proportional to the power dissipation per
unit area JE.
Texturing of the surface aims to reduce the dissipation by
reducing both the current density and the electric gradient,
especially in the core region between the sheds. The proposed
design is a structure which has a textured surface comprising
arrays of protuberances of hyperboloidal, conical and
pyramidal or other symmetrical form. The textured arrays
offer a controlled variation of surface area.
Figure 1 represents protuberances of height c, formed by
domed sections of spheres of radius b, the bases of the
protuberances being circles of radius a.
In this case,
a2 = c (2b - c) (2)
and the surface area of the protuberance is
Ap = 2S b c (3)
These three adjacent protuberances will increase the surface
area of the underlying triangular plane surface of side 2a,
which has a surface area
At = 3 a2 (4)
to a value of
Apt = Ap / 2 + A t - Sa2/2
= a2 [ Sb / (2b-c) + 3 -S/2 ] (5)
Figure 1. Contiguous part-spherical protuberance.
IEEE Transactions on Dielectrics and Electrical Insulation
The surface area is thus increased by the presence of the
spherical protuberances by a factor that is defined by the ratio
Apt / A t = 1 + Sc / [ 2 3 (2b-c )]
The increase is dependent on the choice of the ratio of
protuberance height, c, to spherical radius, b. Hemispherical
protuberances (b=c) will have the limiting value of the area
ratio
[Apt / At]{hemisphere} = 1 + S/2 3 = 1.907
which notably is independent of the radius b of the
protuberance. For such a patterned texture of hemispherical
protuberances in close hexagonal contact, this ratio is close to
the limiting value of 2 given by the ratio of hemispherical and
circular areas. Higher factors can be achieved with other
geometrical forms of the protuberance. For example, a semi-
ellipsoidal protuberance with y=2x has a surface area 3.42S x2,
which gives an increased value of the surface area factor of
3.42 compared with the value of 2 for a hemispherical
protuberance. By controlling surface area in this way, a three-
dimensional patterned texture, suitably dimensioned to
achieve controlled variation of surface area along the
structure, will provide a controlled variation of leakage current
density and surface electric field strength, for a uniform
pollution layer.
E be reduced by the combined power density factor (Į x ȕ).
(a) Texture 1 (b) Texture 2
(c) Texture 3 (d) Texture 4
Figure 2. Some alternatives of possible texture designs.
3.2 CREEPAGE DISTANCE
Conventional insulator design for adverse or polluted
environments does not include this avoidance of dry band
formation by control of surface area. Instead, such anti-fog
designs are based upon increased overall insulator length and
Vol. 17, No. 2; April 2010 345
by profile changes, such as large sheds that reduce the form
factor or a ribbed structure that increases the specific creepage
distance [5]. It is, therefore, important that a textured surface
(6) should, in addition to surface area control, increase the
longitudinal surface (creepage) length of the insulating
structure without then necessarily increasing the overall length
of the structure. The patent [2] describes the use of several
patterned textures of protuberances. This will give an increase
in both surface area and creepage length. Figure 2 shows four
patterns representing plan views of either contiguous or
(7) intersecting protuberances.
The arrowed paths shown in Figure 2 indicate the surface
creepage lengths for these patterns. These are circular arc
paths, although appearing linear in the plan view in Figures
2b, 2c and 2d.
Thus, for example, the longitudinal creepage distance per
row, in the direction of the electric field strength along the
hemispherical patterned texture of Figure 2(a), is 2.303
compared with 3 a on a plain surface. This is a per-unit
creepage distance increase of 1.209. As a result, both E and J
terms of equation (1) are significantly reduced.
3.3 POWER DISSIPATION ON INSULATOR
SURFACE
Suppose that, in any given region of the insulator, the
increased surface area of the texture reduces J by a factor Į.
The texture at the same time also increases the creepage
distance in the field direction by a factor ȕ, and will therefore
reduce the surface electric field strength by the same factor.
The local power dissipation per unit area will, consequently,
The choice of texture will provide independent control of Į
and ȕ. For any proposed texture, we can base the design on the
product Įȕ, which gives a kind of figure of merit of anti-dry-
band performance, so that the higher this figure of merit the
lower the tendency to surface drying. Some examples of
extremes of anti-fog design are first the standard commercial
anti-fog insulator, whose fins, ribs and steps increase only
creepage distance, i.e. the factor (E); and secondly,
longitudinally fluted prototypes that give an increased surface
area factor (D), but no increased creepage distance.
3.4 EROSION REDUCTION
While the primary purpose of the insulator textures is to
inhibit partial arc activity at dry bands, it is equally important
to determine whether discharge activity would lead to
different tracking or erosion than would occur for plane
untextured surfaces. On the contrary, it may be anticipated that
there would be a reduction in tracking and erosion damage for
the textured material. This expectation is based on two
grounds: (i) the increased creepage distance between the
electrodes may reduce the current in the test arc; and (ii) the
arc discharges that cause erosion may be divided by the
texture into less damaging, parallel arcs of lower current. With
this objective, inclined plane tests were performed using an
automated test rig developed at Cardiff.
346 A. Haddad et al.: A New Approach to Anti-fog Design for Polymeric Insulators

3.5 FIGURES OF MERIT For non-contiguous hemispheres with a radius of b and a

It is instructive to compare the power density factors (Įȕ) centre spacing of e, the area ratio of the textured surface to the
associated with a variety of textures that result from an increase plane surface is smaller than that calculated in equation (7),
in both area and creepage factors. Table 1 shows the calculated and becomes
area and creepage factors for the texture types shown in Figure [Apt / At]{non-contiguous} = [2Sb2 + 3 e2] / 3 e2 (8)
2. Thus, on the basis of the results shown in Table 1, Textures 2
and 3 appear to be the most promising for an anti-fog insulator.
Note that all these factors have been calculated on a simple The specimens shown in Figure 3 have a centre spacing of
geometrical basis. So (D) has been derived from the total e=3.4 mm, which gives an area ratio of 1.49.
increase of surface area, and will give an optimistically high
value compared with a non-uniform current density
distribution. On the other hand, (E) will be underestimated by
these calculations of the shortest creepage distance.

Table 1. Area, creepage and power density factors for hemispherical textures
of Figure 2.
Area factor Creepage Power density
Texture
D factor E factor DE
1. Contiguous hexagonal 1.907 1.209 2.306

2. Intersecting hexagonal 1.446 2.356 3.407 (a) Top view

3. Intersecting square 1.301 2.222 2.891

4. Intersecting triangular 1.209 1.814 2.193

4 SAMPLE PREPARATION
Test specimens were moulded from liquid silicone rubber (b) Side view
(RTV-2) Dow Corning HV 1540/10P (Table 2). The
specimens were cured at room temperature for a minimum 24
hours, followed by 3 hours post-curing at 120 °C. Figure 3. Rectangular textured specimen.

The mould was designed to produce rectangular specimens

of dimensions 120 mm x 50 mm and thickness 6 mm. One set 5 INITIAL TEST PROGRAMME
of mould inserts gave specimens with two plane surfaces. A 5.1 INCLINED-PLANE TESTS
second set moulded one side of the specimen with a regular
pattern of hemispheres of diameter 2.5 mm (Figure 2). The Here, the aim was to determine whether the surface tracking
choice of a 2.5 mm diameter was made to provide a possible and erosion properties of the silicone rubber insulation are
additional advantage of a large contact angle in water drop improved by texturing. In the material test specified in IEC
formation on an insulator surface. 60587 [6], the insulation specimen must be inclined with its
mount at an angle of 45q with the test surface on the
Table 2. Properties of silicone rubber HV 1540/20P. underside, so that a specified contaminant flows over the
Property Value surface. Electrodes at a separation of 50 mm, and whose
Dielectric Constant (100 Hz), IEC 60250 2.8 geometry are also specified in the standard, are subjected to a
Dielectric Strength (kV/mm), IEC 60243 24
Dissipation Factor, IEC 60250 0.001 power frequency voltage of 2.5, 3.5, or 4.5 kV rms, depending
Elongation,% 500 on the specification of the material. The standard test criterion
Shore A Hardness, DIN 53505 32 is that the discharge current does not exceed 60 mA during a 6
Solids Content 100 hour test for 2 s, or alternatively, the surface arcing length
Specific Gravity 1.13
Tear Strength (N/mm), ASTM D 624 B 30 does not exceed 25 mm. In the present special tests, textured
Tensile Strength (MPa), DIN 53504 7.8 and plane untextured samples were tested simultaneously.
Tracking Resistance, IEC 60587 1A Three parameters were measured: (a) the maximum rms value
Viscosity Y=10.0 s-1 Pa.s 14 of current recorded during the test, (b) the accumulated energy
Volume Resistivity, IEC 60093 6x1014
dissipation during the full test duration of 6 h, and (c) the
weight loss of the sample as a result of the test. The data
In practice, texturing of the specimens with contiguous acquisition was obtained with a sampling rate of 400
(touching) hemispherical protuberances to give an increase of samples/cycle using an in-house data acquisition program.
surface area 1.907 times that of the plane specimens could not During and after the tests, photographs were taken of the
presently be manufactured in-house, and work is in progress to discharge activity and the ensuing tracking and erosion
achieve this using external sources. damage.
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 17, No. 2; April 2010 347

A test voltage of 3.5 kV rms was applied for 6 hours with a
contaminant flow rate 0.30 ml/min, a concentration of NH4Cl
(ammonium chloride) solution 0.1% and a conductivity of
0.253 S/m at 23 qC. The non-ionic wetting agent Trixton-100
was used at a concentration of 0.02%.
Since the tests are designed to promote arcing, the
specimens in these tests were highly stressed electrically. For
the given electrode distance of 50 mm and the 3.5 kV test
voltage, the specific creepage distance is 14mm/kV
(equivalent to 5.8 mm/kV phase voltage). This is, for example,
three times smaller than the shortest specific creepage distance
allowed in [5] even for a light pollution level. However,
during these tests, transient voltage falls to 2.5 kV occurred,
due to regulation limitations on the high voltage supply
transformer.
5.2 FOG TESTS
The solid-layer clean-fog tests, on plane and textured
rectangular specimens, required the application of an artificial
pollution layer by dipping in or flowing on a contaminant
suspension. A standard suspension [7] was prepared in 1 litre
of water with 100 g of Kieselguhr, 10 g of SiO2 (particle size
< 20 nm) and a suitable quantity of NaCl to achieve a desired
salinity Sa (kg/m3). A guide to the relationship between Sa and
the volume conductivity V20 (S/m) at 20 qC of the suspension
gives
Sa = (5.7 V20)1.03 (9)
The volume conductivity will determine the salt deposit
density (SDD, mg/cm2) on the pollution layer after drying and
the maximum layer conductance (k20, PS) that will arise during
a clean-fog test. Tabulated data [7] suggest the empirical
relationships
V20 = 3 104 k20 (10)
k20 = 400 SDD (11)
where the layer conductivity and volume conductivity, Vl, of
the layer are related by the layer thickness, T,
This integration can be performed using the specimen
profile, where P(l) is the perimetric length of the specimen at a
distance l along the surface creepage path LC. For the
rectangular plane samples, the form factor is unity. For the
textured samples, it is at the same time reduced by the
increased effective surface width, and increased by the
creepage length. Then, the reference layer conductance is
K (LC) = k20F (15)
After mounting the specimen vertically in the fog chamber,
using a purpose built test rig, the reference layer conductance
can be measured by applying a voltage that is proportional to
the creepage length. This must be equivalent to not less than
700 Vrms/m intermittently, so that leakage current can be
measured without drying the layer. This also enables
verification of the fog density, which must achieve a
maximum leakage current IL within a time between 20 and 40
minutes. The reference layer conductance corresponds with
this maximum, and can be calculated from
K (LC) = F u I/V (16)
The choice of K (LC) and SDD for the test depends on the
severity of the environmental conditions for which an
insulator is to be employed. Four levels of pollution are
defined in [5] with four corresponding specific creepage
distances (mm/kV, where this voltage is 3 u rms test
voltage).
These layer conductance and SDD values are experimental
values from test laboratories for several insulator types. It is
noteworthy that these relationships between LS and k suggest
that the leakage current that can be sustained before flashover
is of the order of 1 A.
Table 3. Recommended parameters and sample test data from [5] (Table II)
and [7] (Table B.1) for pollution Class I-IV.
Ranges of values
LS
(mm/kV) Maximum Maximum
withstand withstand SDD
k20 (PS) (mg/cm2)

k20 = Vl T (12)
Light I 16 7 - 14 0.02 - 0.04

The local layer conductivity can be measured before drying Medium II 20 14 - 28 0.04 - 0.1
of the suspension by use of a suitable surface probe. After
drying, the SDD can be checked by removal and collection of Heavy III 25 14 - 40 0.06 - 0.15
the pollution layer from an area A of the specimen and
Very heavy IV 31 28 - 56 -
measuring the conductivity V20 and, hence, the salinity in a
volume V:
SDD = SaV/A (13) The desired test strategy was thus to test at both Class I and
Class IV levels (Table 3); the required SDD and layer
The effective overall conductance (known as the reference conductance were to be estimated from the mid-range of the
layer conductance K(LC)) of the pollution layer on an insulator values shown, taking into account the form factor of the
will depend on its geometrical form factor F: specimens. The volume conductivity of the contaminant
LC
suspension is then determined. For the chosen class of
dl pollution, it is straightforward to calculate the anticipated
F= ³
0
P(l )
(14)
withstand voltages from the table using the product of the
348 A. Haddad et al.: A New Approach to Anti-fog Design for Polymeric Insulators

specific creepage distance and the actual creepage length of pollutant in the inclined-plane test, both the weight loss and
the test specimens. the characteristics of the erosion (Table 5) are significantly
The rating of the test transformers available in the mitigated. This is attributed to the establishment of mobile
laboratory and used in the tests was not sufficient to perform discharge paths, and is expected to improve further for
valid withstand-voltage measurements, which need a short- textures that offer a better approximation to the design
circuit secondary current of about 6 A. Nevertheless, reliable requirements than that of Figure 3.
low-voltage conductance data could be obtained, and at higher
Table 4. IEC 60587 inclined-plane test results.
voltages useful qualitative comparative evaluations were
Test Max. Irms Energy Weight
possible during the pre-flashover phase. A data acquisition sample (mA) dissipation (MJ) loss (g)
program, developed for these pollution tests, calculated many Plane 36 5,2 1,75
parameters as a function of time: voltage rms value, leakage Textured 38 5,1 0,44
current rms value, leakage current absolute peak value, Textured 41 5,0 1,67
leakage current positive peak value, leakage current negative Plane 42 5,5 1,16
peak value, leakage current THD, power factor angle, leakage Plane 32 4,6 1,10
current charge, average power and accumulated energy Textured 34 4,6 0,52
dissipation. The monitor display was divided into 14 windows Textured 36 4,97 0,86
in which the parameter changes were shown. Additionally, the Plane 36 4,99 0,89
shapes of the leakage current and voltage, and the FFT of the Plane 38 4,93 0,78
Textured 35 - 0,30
most recent leakage current cycle were displayed. The
Textured 31 5,15 0,87
program collected data with a speed of 400 samples/cycle for
Textured 42 - 1,50
one voltage channel and two current channels. A test of 100 Plane 40 - 1,04
minutes produced 2.4 G bites of stored data. In this way, Plane 45 5,02 1,58
extensive post-processing of data is possible after the test. Textured 36 5,14 0,92
Plane 40 - 1,63
Plane 43 4,75 0,69
6 TEST RESULTS Textured 35 5,16 0,50
6.1. TRACKING AND EROSION TESTS Plane 28 4,93 1,38
Textured 39 5,23 1,27
a) Absorbed energy and weight loss results
Inclined plane tests were performed according to Method 1 Table 5. Length and depth of erosion.
in IEC60587 [6]. Results from 20 tests are shown in Table 4. Plane samples Textured samples
A positive correlation between the energy dissipated during a
Longest erosion 52 mm 31 mm
test and the consequent weight loss in the specimen is Shortest erosion 13 mm 6 mm
confirmed in Figure 4, for both plane and textured surfaces. Deepest erosion 6 mm, 6 mm,
9 samples 3 samples
Shallowest erosion 5 mm, 3 mm,
b) Partial arcs and erosion observations 1 sample 2 samples
Early in each test, small dynamic discharges appeared,
because the severe voltage stress conditions are designed to 3
Textured
accelerate damage of the test material by arcing. The pollution 2,5
Weight loss ( g )

Plain
severity was also very high. The solution conductivity was 2
0.253 S/m. So, assuming a thickness for the electrolyte layer 1,5
of 0.1mm, a surface conductance of 25 PS led to a very large 1
leakage current. Although during the first minutes of the tests, 0,5
the arcs on the textured samples burned more actively at the 0
lower electrode, after about 10 minutes, a straight arc channel 4 5 6

formed in the case of plane surfaces (Figure 5). Then, whereas Dissipated energy ( MJ )

the discharge on the plane samples damaged one area near the
lower electrode, the discharges on the textured surfaces moved
along the lower electrode. This indicates a multi-path current
conduction which makes the discharge highly mobile and
unable to anchor permanently at any one single point on the
lower electrode. This different behaviour of the arcs can lead
to very large erosion on plane specimens compared with
textured specimens. Table 5 gives some erosion data from
these tests.
Despite the close similarity of the energy dissipation for
both types of sample, suggesting that the leakage current is not
affected by the texture for the very heavy application of
Figure 4. Test specimen weight loss during IEC60587 inclined-plane test.
6.2 DRY-LAYER CLEAN-FOG TESTS
A short series of fog-chamber tests used the same
rectangular plane and textured specimens. A pair of lower
(earthed) electrodes from the inclined-plane test were attached
to the specimen and set at a separation of 50 mm. The
specimens were positioned in the chamber centre so that the
face of the sample was aligned vertically between two round
aluminum plates supported by three polyethylene rods. The
wetting intensity as measured by a special vessel had a
IEEE Transactions on Dielectrics and Electrical Insulation
2
horizontal component of 93 g/m /minute or 0.093 mm/minute,
and a vertical component of 64 g/m2/minute, using tap water
with a volume conductivity 26.7 mS/m at 19 qC.
No flashovers occurred on either plane or textured
specimens at the test voltage of 11.2 kV. However, the
discharge activity on the textured specimens showed marked
differences from that on the plain specimens. On the plane
specimens, the partial arc activity extended from the high-
voltage electrode across 50% or more of the inter-electrode
distance. Subsequent examination revealed that these
discharges had caused this region of the silicone rubber
surface to lose its hydrophobic properties. The partial arcs on
the textured specimens were not anchored to the high-voltage
region, but ignited and extinguished randomly over the surface
between the high-voltage electrode and the earth electrode.
This mobility suggests that any loss of hydrophobicity is
transient at any surface location. Visual inspection could not
discern this. Further work is being undertaken under a follow-
up research project funded by the UK Engineering and
Physical Sciences Research Council (EPSRC).
a) First degradation signs
b) Degradation after 1 hour
Figure 5. Degradation patterns for flat and textured insulation samples.
7 CONCLUSIONS
A new concept of insulating surfaces is introduced. The
design principles have been defined, and initial laboratory
tests were performed on plane and textured rectangular
samples. The results so far achieved are promising in several
respects. It has been shown that tracking and erosion rates are
reduced for textured silicone rubber surfaces. It remains to be
shown to what extent this is due to the subdivision of arcing
Vol. 17, No. 2; April 2010 349
into parallel paths or to the increased creepage length for the
textured surface. The mean layer conductance of the textured
surface appears to be comparable with that of the plane
surface, so that the leakage current is of the same order.
The comparative high-voltage fog tests that have so far
been carried out are also encouraging. The partial arcs
observed with textured surfaces seem to be smaller and more
mobile than with plane surfaces. This may indicate that the
aim of inhibiting the formation of large fixed dry bands is
being achieved.
The specific aim of the ongoing research and development
programme is to produce new silicone rubber insulators that
will offer improved pollution performance at an economic
cost. Apart from the promising preliminary results that have
been presented here, there are further indications of a
successful outcome:
x The performance of a clean hydrophobic silicone rubber
insulator is excellent, with a withstand specific creepage
length (SCL) as low as 6 mm/kV.
However, a moist environment associated with even light
pollution reduces the performance significantly, if at the
same time the pollution layer causes a loss of
hydrophobicity (even if this is temporary). Thus, a layer
conductance of only 2 µS would require a much longer
insulator at about 20 mm/kV. Such a light degree of
pollution is much more common than heavy pollution,
except in specific maritime or industrial locations.
x The insulator leakage current can be significantly
controlled by surface geometry for such light pollution
layers.
x Any large non-uniformity in the surface leakage current
which may arise in textured insulators by localized
heavier pollution or wetting, will be largely and rapidly
eliminated by the greater evaporation rate from areas of
higher conductivity.
x Any locally increased conductance of the textured
insulator surface under fog or light precipitation will
delay dry band formation.
x The economic implications of a large value of the design
SCL of outdoor insulators are large. The production costs
of textured insulators are likely to be increased only by a
small factor associated with the production of moulds.
8 ACKNOWLEDGMENT
The authors would like to thank the Cardiff Partnership
Fund and EPSRC for their financial support.
9 REFERENCES
[1] D.L. Williams, A. Haddad, A.R. Rowlands, H.M. Young and R.T.
Waters,“Formation and characterization of dry bands in clean fog on
polluted insulators”, IEEE Trans. Dielectr. Electr. Insul., Vol. 6, pp.724-
31, 1999.
[2] A. Haddad and R.T. Waters, “Insulating structures”, UK Patent
2406225, 2003.
[3] A. Haddad, H. Griffiths and R.T. Waters, “Principles of anti-fog design
for polymeric insulators”, IEEE Intern. Conf. Solid Dielectr. (ICSD),
Winchester, UK, pp.302- 305, 2007.
350 A. Haddad et al.: A New Approach to Anti-fog Design for Polymeric Insulators
[4] A. Haddad, K. Chrzan, H. Griffiths and R.T. Waters, “A new approach
to anti-fog design for polymeric insulators”, Intern. Sympos. High
Voltage Engineering (ISH), Ljubljana, Slovenia, paper T4-446, 2007.
[5] IEC 60815: 1986: “Guide for the selection of insulators in respect of
polluted conditions”, IEC Standard.
[6] IEC 60587: 1984: “Methods for evaluating resistance to tracking and
erosion of electrical insulating materials used under severe ambient
conditions”, IEC Standard.
[7] IEC 60507: 1991: “artificial pollution tests on high-voltage insulators to
be used on a.c. systems”, IEC Standard.
A. Haddad obtained the degree of Ingénieur d’Etat in electrical
engineering in 1985 and then a Ph.D. in high voltage engineering
in 1990. Following graduation, he took up a Research Associate
position. In 1995, he was appointed a Lecturer and in 2006, a
Professor in electrical engineering at Cardiff University, with
responsibility for the High Voltage Energy Systems Group
(HIVES). His research interests are in overvoltage protection, insulation systems,
insulation coordination and earthing of electrical energy systems. He has co-authored
over 100 publications, with three paper awards. He has recently published an IET-
Power Series Book on Advances in “High Voltage Engineering”. He is a member of
IET and CIGRE. He is chairman of the IET South Wales Power Specialist, and
member of the British Standard Institution committees on overvoltage protection of
low and high voltage systems BSI PEL1 and PEL2, the International Electrotechnical
committees IEC TC37 MT4 and MT10, and a member of the Steering Committee of
the International Universities Power Engineering Conference (IUPEC). He is also a
founding member and current chairman of the UK Universities High Voltage network
(UHVnet).
R. T. Waters received the Ph.D. degree at the University of
Wales, Swansea in 1954. He researched at AEI
Aldermaston, UK with teams on nuclear fusion, high speed
photography, electron microscopy and high voltage
technology. From 1963, his university work at Cardiff
initiated international collaborations in high voltage
engineering and gaseous breakdown, including participation with the
European Les Renardieres Group on UHV phenomena. He has been closely
associated since 1972 with the biennial International Symposium on High
Voltage Engineering of which he is a Steering Committee member. He is a
Fellow of IET (formerly IEE, UK) and is now Emeritus Professor at Cardiff
University.
View publication stats
H. Griffiths joined the UK electricity industry in 1978 as an
engineering apprentice and obtained a BSc. degree from the
Polytechnic of Wales in 1982. Between 1983 and 1990, he
worked at the South Wales Electricity Board and the Central
Electricity Generating Board (CEGB) as an engineer in
distribution and transmission system design. In 1990, he was
appointed to the lecturing staff at Cardiff University, where he obtained his
PhD. He is currently a Senior Lecturer in the HIVES Group. His research
interests include earthing systems and transients. He is currently a member of
BSI PEL/99 and chair of BSI GEL/600. He is a chartered engineer and a
member of IET.
N. Harid graduated in 1985 and then obtained a Ph.D.
degree in electrical engineering in 1991 from the
University of Wales. After several years of academic
experience, Dr Harid joined the HIVES group at Cardiff
in 2001 as a Senior Researcher. He is now a Lecturer at
Cardiff University. His main research interests are in earthing systems and
breakdown phenomena. He is a member of IET and BSI GEL/81.
P. Sarkar (M’06) was born in India, received the
bachelor’s degree in electrical engineering from Jadavpur
University, India. He then joined the Institute for Plasma
Research, India in 1996, working in the Pulsed Power
Group until 2004. In 2008 he received the Ph.D. degree
from Loughborough University, UK and is now a Research
Associate in the HIVES group. His present research interests are overhead line
insulator design and its performance, compact pulsed power devices, and
Tesla transformers.
P. Charalampidis was born in New York, U.S.A.,
received his degree in electrical and computer engineering
from Aristotle University School of Engineering, Greece.
His thesis subject was on the effect of RTV rubber
coatings on flashover withstand of insulators. In 2009, he
joined the School of Engineering at Cardiff University as a
PhD candidate within the HIVES group. His research project is on the
characterisation of textured insulators for overhead lines and substations.