Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

E1 30 PDF

Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

Dimensioning of electrodes in the UHV rangeillustrated with the example of

toroid electrodes for voltage dividers


K. Feser

Dimensioning of electrodes in the UHV range-illustrated with the


example of toroid electrodes for voltage dividers
Kurt Feser

Subject: Electrodes, onset voltage, electric field

Summary: The discussion covers the essential aspects required for


dimensioning electrodes in the UHV range, whereby the breakdown
voltage criterion seems to be purposeful for the generation of lightning impulse voltages at least where large test specimen capacities
are involved. On the other hand, the onset voltage level should be
used as a dimensioning criterion for economic reasons in the case of
high switching impulse voltages (over approx. 2 MV). The permissible surface field strengths for various electrode radii are given for
switching impulse, AC and DC voltages. This data is used to calculate
the maximum permissible voltage for double toroid electrodes.

0.

1. Onset voltage, breakdown voltage


Fig. 3 illustrates the difference between onset voltage Ue and
breakdown voltage Ud for a given electrode configuration.
In this example, onset voltage Ue does not differ from breakdown voltage Ud until a certain distance a E 3 . . 4 m. In zone 1,
the onset voltage and breakdown voltage levels are practically
identical. In zone 2, the breakdown occurs out of stable predischarges. The transition from zone 1 to zone 2 is not abrupt for a certain configuration. It is possible to observe scattering breakdown voltage values in this intermediate zone.

Introduction
kV

Electrode dimensioning for test plants for the megavolt range


is based either on onset voltages or on breakdown voltages.
Previous dimensioning practices, for example for impulse
generators, hardly considered other criteria besides the breakdown voltage level. As long as only lightning impulse voltages
were generated, this dimensioning concept yielded solutions
which were perfectly justified from an economic point of
view (Fig. 1). The clearances to walls or ceilings had to be
chosen in accordance with an electric strengthofabout SOOkV/m.
Apart from the generation of lightning impulse voltages for
testing purposes, however, the generation of switching impulse
voltages with Marx impulse generators has become especially
important for research and development involving 1500 kV
material, since the strength of the air features a distinct minimum in the range of switching impulse voltages. The problem
in the generation of positive switching impulse voltages of
2 MV is the fact that depending on the impulse shape, clearances of 10 to 12 m are necessary between live parts and
earthed objects if breakdowns are to be prevented. The conditions are even more extreme on the 3 MV level (Fig. 1).
Electrode dimensioning for a test plant on the basis of the
breakdown voltage/breakdown distance characteristic is no
longer economically purposeful for voltages over 2 MV, because building costs rapidly rise in view of the large necessary
clearances.

3500
3000
2500
t 2000
Ud
1500

1000
SO0
0

10

15

20

25

30 m 3 5

a-

Figure 1. 50% breakdown voltages of rod/plane spark gaps as a


function of distance a.
4 Lowest positive 50% breakdown impulse voltage for random
impulse shapes
ud = ud

It must also be observed that the pre-discharges which occur


prior to the breakdown influence both the voltage shape and
the voltage withstand capability at the test specimen: for certain test specimens, this is not permissible. The voltage withstand capability is influenced because the pre-discharges can
lead to a change in field conditions. A voltage shape is influenced, because the charge requirement of the pre-discharge can
range up to several ten microcoulomb per meter pre-discharge length (4). This charge requirement leads to a voltage
drop at the excessively high internal resistance ot the voltage
source, thus causing a voltage change at the test specimen itself (Fig. 2).

Figure 2. Influence of the leader pre-discharge on the voltage shape


for switching impulse voltages.

2500

Additionally, literature (7) reports that breakdown probability


in this specific transition zone may deviate from the normal
distribution. Therefore, this zone in which the breakdown
voltage may vary widely is designated as a scattering zone in
Fig. 3 and 4. The deliberations developed in this paper focus
mainly on the lowest breakdown voltage Ud or the nominal
voltage Ue

kV
2000

500
0
m

10

12

1Lm

16

Figure 3. Onset voltage Ue and breakdown voltage Ud of a sphere/


plane spark gap as a function of distance a.
2500
kV

50cm
I,

2000

500
0

a-

10

12

1L m

16

Figure 4. Influence of sphere radius d/2 on onset and breakdown


voltages U, and Ud in a sphere/plane configuration (1).
Stab-Platte-Anordnung = rod/plane spark gap
LO
kV
fi
30

Ilns.rl

"0

For electrode dimensioning purposes, it is very important to


know which parameters influence onset voltages and which
influence breakdown voltages. In the UHV range, two criteria
are of decisive significance for the onset voltage level:
1. the electrode radius and
2. the voltage shape.
In zone 1, the same criteria apply both to the breakdown voltage and to the onset voltage. In zone II, however, the type of
pre-discharge and the polarity of the voltage determined the
breakdown voltage (Fig. 3).
Fig. 4 shows the influence of radius d/2 of the electrode on
the onset or breakdown voltage levels in a sphere/plane configuration for switching impulse voltages. It is apparent that
an increase of the electrode radius will be followed by an
increase in onset voltage. It is also obvious that a larger radius
does not cause a decisive increase in breakdown voltage provided the breakdown occurs out of the same pre-discharge
forms (2). It is also interesting to compare breakdown voltages for the sphere/plane configuration in zone II, i.e. with predischarges with the breakdown voltage of a rod/plane configuration. As soon as the leader pre-discharge occurs with
switching impulse voltages, a difference in breakdown voltages is small; with increasing radius, the breakdown voltage will
increase somewhat (2).
Fig. 5 shows the influence of the voltage shape on the onset
field strength Ee = Ue/(a .n) for a radius of 50 cm (1, 3). An
increase of field strength by about 20% is permissible with
switching impulse voltages in contrast to AC voltages. Furthermore, the onset field strength in the UHV range, i.e. for
large distances, is practically independent of clearance values.

12

a-

lm8

Figure 5. Onset field strength E, of a sphere/plane configuration ,I \


as a function of distance a with the voltage shape as the parameter <i j.

In zone II mentioned above, i.e. when dimensioning on the


basis of the breakdown voltage with previously occurring
pre-discharges, the type of pre-discharge is decisive. The socalled leader pre-discharge is responsible for the substantial
reduction of the breakdown voltage in conjunction with
switching impulse voltages. This leader pre-discharge was
previously observed with switching impulse and AC voltages
(5). In the pre-discharge channel of the leader pre-discharge,
a voltage drop of 50 to 200 kV/m is anticipated so that at
3 MV, breakdowns over a distance of several 10 m may be
expected. Literature often mentioned abnormal breakdowns probably because these breakdowns do not follow the
law of the shortest flashover distance whatsoever.

2. Dimensioning of toroid electrodes


After discussing these general electrode dimensioning criteria,
we shall investigate the interesting example of toroid electrodes for voltage dividers in the UHV range. Fig. 6 shows the
layout of a voltage divider. In the following deliberations,
we shall mainly vary the dimensions of the top electrode
(D, d). The height H = 14.5 m corresponds with a voltage
divider for a lightning impulse voltage of about 5 MV.
$1(1
Figure 6. Toroid electrode dimensions for a voltage divider.
4

If a double toroid with identically dimensioned rings - i.e.


dl = d2 - is selected as a top electrode for this voltage divider
as this frequently occurs. Fig. 7 indicates the computer-cal-

culated surface field strengh E at both rings for various ring


diameters (dl/2)/(d2/2) (3). The specified field strength
values apply in conjunction with a voltage U of 1 MV and of
course are applicable only as long as no pre-discharges occur,
The field controlling effect of the capacitors is considered in
the calculations insofar as a linear voltage distribution was
assumed to exist along the voltage divider. For 3 MV, for
example, a double toroid electrode with a ring radius of (dl/
2) / (d2/2) = 60 cm/60 cm would produce a field strength
of 3*5,7 kV/cm at the bottom electrode ring. It is apparent
that the field strengths at both rings increase with decreasing
radius. For practical purposes, it is interesting to know that
the field strength at the lower ring is smaller than at the
upper ring. In optimization calculations, this means that the
size of the bottom ring can be reduced. For radius reductions in intervals of 10 cm of the lower ring, Fig. 8 shows
that the maximum surface field strengths of both toroids
become similar.

Ema,

Figure 7. Maximum field strengths Em, at identical rings - dl =


d2 - of double toroid electrode for a voltage U = 1 MV.

18

kV
on

16

The maximum field strength Em, for double toroid electrodes is plotted over an equivalent radius Pe in Fig. 9 as
determined with the aid of toroid dimensions&I/d and h/d
according to Thione (11) (Fig. 10). The applicable equation
is:
Peq = dylm
For a sphere is peq = d/2 and for an mdividual ring peq = d/m.
For larger toroid electrodes - for example for d/2 = R =
60 cm - a specific problem is the fact that the toroid diameter D will be about 5 m, thus permitting the field strength
in the electrode to attain substantial values. This means that
large radii must be used inside the electrodes or that further
toroids (triple to quadruple toroids) must be installed. Special
attention must be directed to the electrode mounting components.

12

Em 1 o

Figure 8. Maximum field strengths Em, at different sized rings


- dl = d2 - 20 cm - of double toroid electrode for a voltage
U=lMV.

0.7tit
m
03

12

5
0

50

100
oeq--

LI
II
- 20 20
303010 20 20 30

-_

I
150

cm

g&3
so*' 50*+150 60
30 LO
El
CiJTlo
IdI L'I/ld~/21-

200

60
m

Figure 9. Maximum field strengths Em, plotted against the equivalent radius peq for a voltage of U = 1 MV. For (dl/2) / (d2/2*) is
applicable for Dl/D2 = 350/230 and**) for Dl/D2 = 450/350.

10 20
Llld-

50 100

Figure 10. Dimensions required to determine the equivalent radius


peq of double toroid electrodes and simple rings.
a) electrode forms

b) m = f (D/d)

hid

D/d

10

1,17

1,25

1,25

20

10

20

1,25

1,lO

1,15

1,15

1,15

For dimenslonmg purposes, the maxmum permissible voltage at which no pre-discharges yet occur for a certain electrode configuration can be determined on the basis of the
maximum permissible field strength Eeper. which is measured
and calculated for various radii (dl/2) / (d2/2) and of the
computed surface field strength. In Fig. 11. the maximum
permlsslble voltages Umax (curve 1) is determined for a
switching impulse voltage 250/2500 and compared with
experimentally measured points. It is apparent that a surface
field strength Eper ( curve 2) of about 22 kV/cm is permissible
for large electrode radii (1). For a double toroid 60/60, for
example, this would yield a onset voltage of approx. 3.82 MV.
Since the permissible onset field strength Eeper. (Fig. 12)
depends on the voltage shape, different onset voltages Ue are
obtained for DC, AC and switching impulse voltages (Fig. 13).

21
Urnox

1
3
A
00

25

50

100
125
@es -

75

0
II

I/I/

-20203030
- - 10 20 2030

150

175

III,

@I &?
fl
yo5J6J
30 LO
LO
L O 50 50
ld~/Zl/id~/21 -

200 cm 2250
I

For large electrode radii, the permissible field strengths are in


the order of 22 kV/cm for switching impulse voltages, of
15 kV/cm for AC voltages and of 13 kV/cm for DC voltages
(1) (Fig. 12). Hence, a double toroid 60/60 is suitable for
DC voltages of 2.25 MV, for AC peak voltages of 2.7 MV and
for switching impulse voltages of 3.82 MV (Fig. 13). Table 1
lists the onset voltages Un for various toroids. The values
apply to clean and smooth, electrodes.

&l
60

Figure 11. Permissible onset field strength Eeper. and maximum


permissible switching impulse voltage Urn, in au plotted against
the equivalent radius peq.
Eezul = Eeper.
l Umax

3EforU=lMV

Eper

115
kV
cm

3. Practical conclusions

35

v\

30

I\

The results obtained establish a method for plotting the


correlation between the necessary clearance and the necessary radius for various onset voltages. This possibility is
interesting because it is often economically purposeful to
select a larger electrode to achieve a certain desired voltage
while simultaneously reducing the laboratory dimensions.

25

t
Eezu~

20
15

I
0

I
25

I
50

I
75

100

125

150

175

200 cm 225

t%-

Id#l/ldzRI -

oram

Figure 12. Permissible onset field strength Eeper. for various voltage shapes plotted against the equivalent radius peq
1 For switching impulse voltage 3 For DC voltages
3 250/2500
Emax for 1 MV and H = 14.5 m
2 For AC voltages at 50 Hz
Table 1. Onset voltage U, for various toroid electrodes with a
distance H = 14.5 m.
Dimensions:
d1/2
cm
20
20
30
30
40
40
50
50
50
60
60

This correlation is shown in Fig. 14 for various switching


impulse voltage amplitudes. With an insufficiently large
electrode, for example, a minimum clearance of 10.5 m
would be necessary to generate a switching impulse voltage of 2 MV without a breakdown. For larger double toroid
electrode radii over about 30/30, the necessary distance to
earthed objects is substantially reduced and is only about
2.1 m for a toroid 60/60.

D1
cm

150 10
150 20
160 20
160 30
230 30
230 40
350 40
450 40
450 50
500 50
500 60

2,s
20

Onset voltages Ue

toroids
&I2
cm

t
:! e

9
cm
140
150
150
160
160
230
230
350
450
450
500

h
cm
70
80
1 0 0
120
140
170
210
240
300
300
320

25012500
MV
1,82
1,9
2,l
2,15
2,48
2,6
2,9
3,15
3,35
3,58
3,82

AC voltage DC voltage
MV
MV
1,45
135

1,6

1,62
1,82

199
2,1

2,25
2,4
2,5

297

12

1,25
1,32
1,35
1,52
1,58
1,75
1,88

LO
2,1

2,25

15
1.0
50

I
75

I
100

I
I
125
15P
e ep-

175

I
I
200 c m 225

Figure 13. Onset voltage U, for various voltage shapes plotted


against the equivalent radius @eq.
1 For switching impulse voltage 250/2500
2 For AC voltages at 50 Hz
3 For DC voltages

This figure clearly shows that it is economically significant


for high voltages over approx. 2.5 MV to dimension the electrodes toobig in order to permit small distances to walls or
ceilings.
Fig. 15 shows screening electrode dimensions plotted against
laboratory sizes for a switching impulse voltage of 2.5 MV. It
is very clearly visible that enormous laboratory dimensions
would be necessary if the electrode of a voltage divider is too
small. If a small voltage divider electrode is chosen for a
switching impulse voltage of 2.5 MV, for example, this
would necessitate a building height of almost 34 m (Fig. 15a);
given optimum electrode dimensioning, however, this height
can be reduced to 17 m (Fig. 15b). The fact that this significant advantage cannot always be utilized for practical purposes is based on the necessity that all live objects such as
supply lines, test specimens, etc. must be considered in this
concept.
Onset voltage-based dimensioning is particularly important
for switching impulse voltages. If a laboratory of a certain

32
m

size is available, economic factors will usually be the reason


for operating with minimum electrode. Dimensioning of
electrodes for test plants for production still occurs on the
basis of lightning impulse voltages, because switching impulse voltage testing has not been generally introduced.
This paper only discussed double toroids. The permissible
field strength can also be applied for other electrode shapes
dimensioned in accordance with the relevant equivalent
radius.

4. Literature
111
121
I31
I41
[51

I61

L. Thione: A study of the dielectric behaviour of high voltage


electrodes in view of their indoor and outdoor auolication.
_ _
CIGRE-SC 33-73 (WG 03).
R. E. Wootton; A. H. Cookson:

Switching impulse breakdown


probability distributions for 5 and 6 meter s here-to-plane gaps
in air. Third Internat. Conf. Gas discharges, eondon 1974.
Conference Publ. Nr. 118, S. 426-429.
H. Steinbigler: Digitale Berechnung elektrischer Felder.
ETZ-A Bd. 90 (1969) S. 663-666.
F. Heilbronner: Dber das Durchschlagverhalten langer Luftfunkenstrecken unter positiver Schaltstogspannung.
Bull.
Schweiz. elektrotechn. Ver. Bd. 65 (1974) S. 491-495.
K. Feser: Mechanismus zur ErkMrung des Schaltspannungsphlnomens. STZ Bd. 46 (1971) S. 937-946.
K. Feser: Probleme bei der Erzeugung hoher SchaItstoL)spannungen im Priiffeld. Bull. Schweiz. elektrotechn. Ver. Bd. 65
(1974) S. 496-508.

(71

C. Menemenlis; G. Harbec: Coefficient of variation of the


positive-impulse breakdown of long air gaps. IEEE Trans. on
Pow. App. & Syst. Bd. 93 (1974) S. 916-921.

Manuscript received 1975 - 01 - 06

Bild 14. Minimal notwendige Hallenabmessungen bei Schaltstohspannungen 250 I2500 in Abhangigkeit vom Ersatzradius Pen.
I 3,s MV
2 3MV
3 '2,5 MV
4 2-h

Bild 15. Gr68e von Laboratorien bei 2,5 MV SchaltstoDspannung.


a) fti kleine Elektrodendurchmesser von 0,2 m Dmr.,
b) fti optimale Elektrodendurchmesser von 1 und 0,8 m Dmr.

You might also like