Paper presented at ISH 2013, August 29, Seoul, Korea
SWITCHING IMPULSE TEST OF LARGE SPHERE-PLANE AIR-GAPS
WITH PROTRUSION ON LARGE SPHERES
1*
2
2
Dong Wu , Liliana Arevalo, Ming Li and Mats Larsson
1
ABB HVDC, SE-771 80 Ludvika, Sweden
2
ABB Corporate Research, SE-721 78 Västerås, Sweden
*Email: [email protected]
Abstract: To improve the insulation strength, larger electrodes, often in the shape of
sphere and toroid, are used. In practical applications, the surface of such a large
electrode may often be “damaged” by holes or protrusions, intentionally or unintentionally.
In order to take such effects into the consideration during the design, there is a need to
study the “damaged” electrodes. This paper presents the results from the laboratory study
on the effects of protrusions on the dielectric strength of sphere-plane gaps. The test
results have confirmed the drastic reduction of dielectric strength by the installation of a
protrusion on large spheres. The levels of reduction depend on the lengths and locations
of the protrusions on the surface of the spheres. However, in all the cases studied, the
U50 values obtained were higher than that of the rod-plane gaps.
1
INTRODUCTION
As the voltage level of HVDC system is increased,
so is the switching impulse voltage level. Switching
impulse voltage is often the dimensioning
parameter for the external insulation design at
indoor condition. To improve the dielectric strength,
larger electrodes, often in the shape of sphere and
toroid, are used. In practical applications, however,
the surface of a large electrode may often be
“damaged” by holes or protrusions, intentionally or
unintentionally. In order to take into consideration
in the design there is a need to study the
“damaged” electrodes.
which
the
high-voltage
electrodes
were
suspended. All the test results presented here
have been corrected to the standard reference
atmosphere. The correction was made according
to the procedure given in IEC60060-1.
2.2
Set-up of the rod-plane gap
A rod-plane gap was tested with 6 meters gap
distance. The rod used was an aluminium rod of 3
meters long and 30 mm in diameter. The rod was
terminated on top with a double-toroid structure,
with 1 meter outer diameter, as shown in figure 1.
Although it is know to us that such an electrode
with high field stress on the surface will be
sensitive to surface irregularities [1-3], studies with
even larger electrodes and at higher voltage are
necessary. In this paper, the first step of a more
systematic study of the damaged electrode is
reported.
2
2.1
EXPERIMENTAL
General
All the tests reported in this paper were performed
with switching impulse of positive polarity. The
waveform was 275/2500 s. The voltage level of
50% breakdown probability, U50, was obtained for
each test set-up by the well known up-and-down
procedure with 30 valid voltage applications in
each test. During the test, the applied voltage and
the waveform of the voltage were recoded. Two
digital cameras were used to record the trajectories
of the discharges.
In all cases, the test objects, rod-plane or sphereplane gaps, were vertically installed with a distance
of 12.8 meters to the nearest wall of the laboratory.
The gap distance was adjusted by a hoist under
Figure 1: The set-up of the rod-plane gap
2.3
Set-up of the sphere-plane gaps
The spheres of different diameters, 1.3, 1.6 and
2.0 meters were used in the test. They are made of
aluminium. The sphere was connected with an
aluminium tube of 3 meters in length and 450 mm
in diameter. Above the tube, a large double-toroid
structure, with outer diameter of 2.5 meters, was
used to terminate the tube. With such a test set-up,
as shown in figure 2, all breakdowns occurred
during the tests were from the spheres.
Figure 4: Different locations of the protrusion on
the sphere, i.e. at the under side or right side of
the sphere. In these photos the discharges were
initiated from the protrusions.
3
Figure 2: The set-up of a sphere in the sphereplane gap
2.4
Protrusions on the spheres
A protrusion is made of a steel rod of 16 mm in
diameter with a half-spherical head. It was so
mounted that it protruded vertically from the
surface of the sphere. Two different lengths, 10
and 50 mm, were tested.
TEST RESULTS
3.1
With a rod-plane gap
The test on a rod-plane gap was performed as a
reference test for the whole test set-up. For testing
with standard impulse waveform, 250/2500 s, the
U50 of a rod-plane gap can be evaluated by the
well know Paris formula [4]:
U50RP = 500 d
0.6
(1)
where: d = gap distance in meter
U50RP = 50% breakdown voltage in kV
The U50 obtained from the test for this rod-plane
gap of 6 m was 1480 kV. This value is 1% higher
than what can be evaluated by the Paris formula.
3.2
With sphere-plane gaps
The sphere-plane gap with a sphere of diameter
1.3 m was tested at two gap distances, 6 and 8 m.
The sphere-plane gap with a sphere of 1.6 m was
tested at one distance of 4 m. The test results are
given in table 1 below. All breakdowns were from
the surface of the sphere.
Table 1: Test results with sphere-plane gaps
Figure 3: A rod protrusion of 10 mm in length on a
sphere of 1.6 meter in diameter
A protrusion was installed either at the under side
of the sphere (i.e., at a position on the vertical axis
of the sphere-plane gap) or at the right side of the
sphere (i.e., at a position 90 degree from the axis).
For each test, only one protrusion was installed at
one location with a given length. The locations of
the protrusion are more clearly shown in figure 4.
The photos in figure 4 were taken when discharges
took place from the protrusions.
Sphere
diameter (m)
Gap distances
(m)
U50
(kV)
Std. Dev.
(%)
1.3
1.3
1.6
6
8
4
2227
2331
2381
2.4
2.7
2.4
3.3
Sphere-plane gaps with protrusions
For the sphere of 1.3 m in diameter, tests with
protrusions were performed at one gap distance, 8
m. The test results are given in table 2.
Table 2: Test results for 1.3 m sphere with
protrusions and with the gap distance of 8 m
Protrusion
Lengths
(mm)
10
50
Locations of the protrusion
Right side
Under side
U50
Std. Dev.
U50
Std. Dev.
(kV)
(%)
(kV)
(%)
2059
1856
5.9
4.1
2153
1861
4.9
5.6
In all the above tests, all breakdowns were from
the protrusions.
4
DISCUSSIONS
It is clearly shown through the tests that a small
protrusion reduces drastically the U50 value of the
sphere-plane gap; as given in figure 5 for the test
results on the 1.6 m sphere.
For the sphere of 1.6 m in diameter, test with
protrusions were performed a three gap distances,
4, 6, and 8 m. The test results are given in table 3,
4, and 5 with respectively.
Table 3: Test results for 1.6 m sphere with
protrusions and with the gap distance of 4 m
Protrusion
Lengths
(mm)
10
50
Locations of the protrusion
Right side
Under side
U50
Std. Dev.
U50
Std. Dev.
(kV)
(%)
(kV)
(%)
1602
1432
6.9
3.7
1484
1248
5.8
5.3
Table 4: Test results for 1.6 m sphere with
protrusions and with the gap distance of 6 m
Protrusion
Lengths
(mm)
10
50
Locations of the protrusion
Right side
Under side
U50
Std. Dev.
U50
Std. Dev.
(kV)
(%)
(kV)
(%)
1900
1657
6.3
3.8
1955
1561
5.4
5.3
Table 5: Test results for 1.6 m sphere with
protrusions and with the gap distance of 8 m
Protrusion
Lengths
(mm)
10
50
Locations of the protrusion
Right side
Under side
U50
Std. Dev.
U50
Std. Dev.
(kV)
(%)
(kV)
(%)
2103
1941
3.5
4.6
2202
1812
5.5
4.7
For a sphere of 2.0 m in diameter, test with
protrusions were performed with on gap distance,
4 m. The test results are given in table 6.
Table 6: Test results for 2.0 m sphere with
protrusions and with the gap distance of 4 m
Protrusion
Lengths
(mm)
10
50
Locations of the protrusion
Right side
Under side
U50
Std. Dev.
U50
Std. Dev.
(kV)
(%)
(kV)
(%)
1659
1514
5.6
4.4
1536
1295
6.1
8.0
Figure 5: Test results with the 1.6 m sphere:  =
sphere-plane without protrusion; x = protrusions
located on the right side; + = protrusions located
on the under side;  = rod-plane gap according to
Paris formula
A longer protrusion, 50 mm, will cause more
severe damage than a shorter protrusion. For the
longer protrusion, the level of the damage is more
severe if the protrusion is located at the under side
than at the right side. However, for the shorter
protrusions, the effects of the locations become
less obvious.
In figure 5, at a longer gap distance, e.g., 8 m, a 10
mm protrusion on the right side led to a lower U50
value than that at the under side. The same trend
can be observed on the 1.3 meter sphere at 8 m
gap distance, as given in table 7. Different from the
results for the 1.6 m sphere, for the 1.3 m sphere,
even with the longer protrusion, 50 mm, the effects
of the locations are not significant.
Table 7: Comparison of the U50 (kV) values for 1.3
and 1.6 m spheres with protrusions and at the gap
distance of 8 m
Sphere
diameters
(m)
1.3
1.6
Locations & lengths of the protrusions
10 mm
50 mm
Right
Under
Right
Under
2059
2103
2153
2202
1856
1941
1861
1812
At a shorter gap distance, e.g. 4 m, the effects of
the locations are significant for both the 10 and 50
mm protrusions, as shown in table 8, for both 1.6
and 2.0 m sphere. A protrusion at under side
results in more severe damage to the sphere.
Table 8: Comparison of the U50 (kV) values for 1.6
and 2.0 m spheres with protrusions and at the gap
distance of 4 meters
Sphere
diameter
(m)
1.6
2.0
Location & length of the protrusion
10 mm
50 mm
Right
Under
Right
Under
1602
1659
1484
1536
1432
1514
1248
1295
With a protrusion installed, a larger standard
deviation than that without a protrusion has been
observed in the test results. This may be explained
by the local high field contributed by the protrusion.
This local high field has modified the discharge
process of the sphere-plane gaps and increased
the spreading of the discharge probability
distribution.
In real design, there are many different electrode
structures with large electrode damaged in
different levels. The test results presented in this
paper represents only a very specific situation.
More studies on this aspect are warranted.
5
The surface electric field is more uniformly
distribution for a smaller sphere at a longer
distance to ground than that for a larger sphere at
a shorter distance to ground. The effects of the
protrusion with respect to its location are therefore
different accordingly.
Although the installation of a protrusion resulted in
a drastic reduction of the U50 in comparison to the
original sphere-plane gaps, for all the cases tested,
the U50 values obtained were still higher than that
of the rod-plane gaps, as shown in figure 6. On
one hand, a gap factor of 1.3 may still be achieved
with a damaged sphere of 1.6 m. One the other
hand a near rod-plane gap situation may be
reached even when a 2.0 meter sphere is
damaged. For the conditions with a 50 mm
protrusion installed at the under side of the
spheres, the impact of the sphere diameters have
limited impact on the U50 values.
CONCLUSIONS
The effects of protrusions on the dielectric strength
of sphere-plane gaps have been studied by
laboratory tests. The test results have confirmed
the drastic reduction of dielectric strength by the
installation of a protrusion on large spheres. The
levels of reduction depend on the lengths and
locations of the protrusions on the surface of the
spheres. However, in all the cases studied, the U50
values obtained were higher than that of the rodplane gaps.
The results presented in this paper are only a small
part of the complicated situation of damaged
electrodes. More studies on this aspect are
warranted.
6
REFERENCES
[1] H. M. Schneider, F. J. Turner: “Switching-surge
flashover characteristics of long sphere-plane
gaps for UHV station design”, IEEE TPAS, Vol.
PAS-94, No. 2, pp. 551- 559, March/April 1975
[2] F. A. M. Rizk: Effect of large electrodes on
sparkover characteristics of air gaps and
station insulators” IEEE TPAS, Vol. PAS-97,
No. 4, pp. 1224- 1231, July/Aug 1978
[3] C. Menemenlis, G. Harbec, J. F. Grenon:
“Switching-impulse corona inception and
breakdown of large high-voltage electrodes in
air” IEEE TPAS, Vol. PAS-97, No. 6, pp. 23672373, Nov/Dec 1978
[4] L. Paris, R. Cortina: “Switching and lightning
impulse discharge characteristics of large air
gaps and long insulating strings” IEEE TPAS,
Vol. PAS-87, No. 4, pp. 947-957, April 1968
Figure 6: The ratio of U50 values obtained in tests
to the U50RP evaluated by Paris formula:  = 1.3 m
sphere-plane without protrusion; x & + = 1.6 m
sphere with protrusions; x & + = 1.3 and 2.0 m
spheres with protrusions.