H.M. Ryan
1. 1 Introduction
This chapter introduces the subject of gaseous insulation
and provides information relating to the application of
gaseous insulants to high voltage systems. It examines
atmospheric air, compressed gases and illustrates how, by
linking available experimental test data from such sources
with a knowledge of the 'effectiveness' of various practical
gas-gap clearances, the designer can achieve reliable
insulation design. The chapter also briefly discusses the
need for high voltage and extra high voltage (EHV) test
areas or laboratories. Evidence is presented of how
laboratory studies, on representative insulation systems and
electrode arrangements provide the designer with choices
relating to electrical stresses, clearance levels, service
performance and testing procedures. Gas insulated
substations (GIS) using sulphur hexafluoride (SF 6) gaseous
insulation have been used in transmission systems
worldwide for more than 35 years. The service reliability of
this class of equipment is of paramount importance. In
addition, the chapter presents a large amount of
experimental breakdown information on SF 6 and briefly
reviews the application of field computation strategies in
support of GIS and other equipment designs. Several of the
major factors influencing the insulation design and in-service
behaviour and reliability of SF6 gaseous and epoxy resin
support insulations, as used in GIS equipment, will also be
considered.
Atmospheric air is the most abundant dielectric material
which has played a vital role in providing a basic insulating
function in almost all electrical components and equipment.
1
Figure 1.1 50% flashover voltages for (a) rod-rod and (b)
rod-plane gaps
Figure 1.2
11
Figure 1.6
14
Figure 1.8
Typical
surface gradients on overhead transmission line
conductors see EHV transmission line reference book, Edison
Electric Inst., New York, 1968
Figure 1.9
field breakvoltages for
mixtures
Uniformdown
SF6 air
18
Figure 1.10 Pashen curves for SF6 N2 mixtures at 20oC (50 Hz)
19
20
(see Figure 1.11) of N2/SF6 mixtures for gas insulated substations (GIS). Similarly, present day difficulties during
planning/ commissioning of certain new overhead line(s),
due partly to delays caused by consideration of a plethora of
'alleged' environmental concerns raised by small protest
groups, has resulted in serious consideration being given to
the adoption of an alternative line technology, the so-called
gas insulated line (GIL) approach. N2/SF6 gas mixtures have
again been suggested and evaluation studies are currently
being carried out by EDF and others in Europe.
Ternary gaseous dielectrics have been extensively
investigated, such as N2 + SF6 or CHF3 + SF6, in conjunction
with perfluorocarbons. Fluoro-carbons, when mixed with SF 6,
may exhibit interesting synergistic effects which at present
are not fully explained. A variety of gaseous insulating
systems has been investigated (see CIGRE Report 163,
referred to above), comprising multi-component gas
mixtures carefully selected on the basis of physicochemical
knowledge, especially on the interactions of low energy
electrons with atoms and molecules. Much valuable research
work in this field has been carried out by a group, led for
many years by L.G. Christophorou, at Oak Ridge National
Laboratory, Knoxville, TN, USA.
1.4 Switchgear and GIS
1.4.1 Introduction
High voltage switchgear, which forms an integral part of any
substation (i.e. switching station), is essentially a
combination of switching and measuring devices [2, 17]. The
switches (circuit-breakers), connect and disconnect the
circuits and the measuring devices (instrument transformers), monitor the system and detect faults. Particular
care must be given to the selection of switchgear since
security of power supply is dependent on their reliability. The
reliability of any circuit-breaker depends on insulation
22
oil -
25
Very
high
unit interrupting ratings can now be achieved with both
systems, though the high performance single-pressure
system requires a more powerful operating mechanism to
ensure a suffi-ciently high velocity for the contact/ cylinder
arrangement. This, however, is easily obtainable with either
pneumatic or hydraulic power units, and at the lower
performance levels, spring opening mechanisms also offer a
practical alternative.
Initially the performance limitations of the single-pressure
system meant that earlier types of circuit-breaker incorporated double-pressure interrupters; but the continuing
development of the single-pressure puffer interrupter has
led to designs which are capable of the highest ratings. It
seems likely, therefore, that such interrupter systems will be
the basis of most future EHV circuit-breakers. They will also
find wider application in the high voltage distribution field.
In any axial gas blast interrupter, the unit performance
may be significantly improved by using a dual-blast
construction in which the arc is drawn through two nozzles.
In this construction, the gas now is towards both arc roots,
and metal vapour from the contacts is not blown into the arc
column. In the partial duo-blast construction (Figure D) one
of these nozzles has a smaller diameter, which reduces the
amount of gas passed through the nozzles and hence also
the mechanical energy input without a significant reduction
in performance compared to the full duo-nozzle
construction. The Reyrolle puffer interrupters incorporate
26
data
Figure 1.14 shows a selection of typical SF 6 breakdown
characteristics for the electrode systems illustrated in Table
1.3 [13]. From the vast amount of breakdown data accrued,
the limits of lightning of large practical GIS gas-gap type
configurations were produced and are illustrated in Figure
1.15 for SF6 pressures in the range (0.1 < p < 0.6) MPa.
Figure 1.13
of 420 kV
equipment
recent
project in
East
VA Tech
Example
GIS
on a
substation
the Middle
(Courtesy
Reyrolle)
29
30
.*
Figure
1.15
Typical limits of 50% breakdown gradient (E50) and critical
withstand gradient (Ew) on SF6 pressure for large coaxial and
perturbed cylindrical electrode systems under clean conditions
[13] curves
(a),(b),(c),(d): Lightning impulse (LI) waveshape (1.2/50s)
(e),(f),(g),(h):
Switching
impulse
(SI)
waveshape
(250/2500s)
also show in curve c, the lower limiting 50 Hz withstand
characteristic (Ew)
E50 data: curves a, b, e and f; Ew data: curves c, d, g and h
33
36
Figure
1.19
0.45 MPa)
power
frequency
(V/t)
39
Figure
1.22 Normalized 50 Hz service ELP levels for SF6 insulated
switchgear equipment.
a, A
b, B
e,c,C
f,d,D
a,b,c,d,e
A,B,C,D
pressure
Working Pressure:
a, A
0.38 MPa
b, B 0.38 MPa
c, C 0.40 MPa
d,D 0.55 MPa
Critical areas of metalclad designs have been identified
which merit special manufacturing, testing and assembly
controls, and these aspects will be further considered in
Section 1.5 and also in later chapters. Several papers have
considered the power frequency (V/t) characteristics of
gaseous and solid insulation using model gas-gaps and
gas/insulator arrangement for large electrode systems under
conditions with gross contamination present and with
varying gas pressures [13].
1.4.4 Performance under contaminated conditions
As has been demonstrated above, the achievable design
stress and reliability of SF6 insulated apparatus under normal
41
life expectancy,
where is the potential at any point in the Cartesian coordinate system x, y, z. The field distribution in any design is
dependent on the shape, size and disposition of the
electrodes and insulation and also, in general, on the
permittivities of the insulating materials used.
One criterion which can be applied to the evaluation of
the electrostatic design of high voltage equipment is the
ratio of the average to maximum field. This parameter,
termed the utilisation factor , is considered below, together
with details of a related parameter termed the normalised
effective electrode separation. Brief reference is made in
Section 1.5.1.3 to a simple approximate two-dimensional
method which can often be used to estimate maximum
voltage gradients for complicated arrangements which
cannot be solved directly by available numerical methods.
Analytical solutions of the Laplace equation can only be
obtained for relatively simple electrode systems, where the
conducting surfaces (electrodes) are cylinders, spheres,
spheroids or other surfaces conforming to equipotentials
surrounding some simple charge distribution. Generally, the
multiplicity of boundary conditions for the complicated
contours encountered in high voltage equipment means that
analytical solutions of the potential are not possible.
Because of this difficulty, several approximation methods
have been investigated, the more important of these being
(i) analogue methods and (ii) numerical methods [21, 34].
46
1.5.1.1
47
2. hyperboloid (points)
3. hyperboloid (point)-plane
value p, for a three-dimensional geometry is lower than
that for the corresponding two-dimensional geometry (e.g.
compare cylinder-plane and sphere-plane, curves 2 and 5,
Figure 1.24). Once again, this may be attributed to the
greater
field
divergence
in
the
three-dimensional
arrangements. Brief mention will be made later to the simple
relationship existing between for corresponding two- and
three-dimensional systems. To summarise, utilisation factors,
or field factors, or field factors, may be calculated for many
practical arrange-ments either as a result of precise analysis
48
Figure1.25
Dependenccd of X upon p for several standard
geometries
Curve I parallel cylinders; curve 2 cylinder-plane; curve 3
concentric cylinders; curve 4 sphere-sphere (symmetrical
supply): curve 5 sphere-sphere (unsymmetrical supply); curve
6 sphere -plane; curve 7 concentric spheres.
revolution
(b) Applies for systems shown in Figure 1.25
(a) Part of single phase layout; (b) field plot at lower end of
bushing; (c) field plot of barrier supporting isolator, (d) field
52
3.
4.
Since the work [34, 37, 41] was carried out, much
valuable information has appeared in the literature
regarding finite element, finite difference, boundary integral
and other variants. Numerous commercial field analysis
packages are now available. Indeed, further examples are
presented in later chapters. A detailed discussion of relation
methods, difference formulas, accelerating factors, etc. is
outside the scope of this chapter. A strategic message from
this writer is that, whatever package is to be used, care
should be exercised to confirm its effectiveness, and the
achievable accuracy, by means of simple validation studies.
53
54
rE = K1pr + K2
(1.2)
where K1 and K2 are constants, p = [(r + g)/r] and r is the
radius of the inner cylinder or sphere. The real virtue of this
simple empirical breakdown estimation technique is the fact
that one can often consider practical electrode systems to
be perturbations of the same simple geometry [37].
GAP g,
mm
Experimental
Empirical*
Semi-empirical
Experimental
Empirical*
Semi-empirical
20
100
102
103
158
158
162
210
214
220
40
127
126
129
200
197
203
259
267
276
59
Experimental
Empirical*
Semi-empirical
144
143
142
80
227
222
227
288
301
310
60
21
22
23
24
25
26
27 ISHIKAWA,
M.
and
HATTIRI,
T.:
'Voltage-time
characteristics of particle initiated breakdown in SF 6 gas'.
3rd Int. Symp. On Gaseous dielectrics Knoxville, USA,
1982, Paper No. 28
28 ETEIBA,
M.G.
and
R1ZK,
A.M.:
'Voltage-time
characteristics of particle- initiated impulse breakdown in
SF6, and SF6,-N2, IEEE Trans. Power Appar. Syst., 1983,
PAS-102, (5), pp. 1352-1360
29 TAKUMA, T. et al: interfacial insulation characteristics in
gas mixtures as alternative to SF6 and application to
power equipment', C1GRE, 2000, Paper 15-207
30 HAMPTON, B.F., PEARSON, IS., JONES, C.I, IRW1N. T.,
WELCH, I.M. and PRYOR, B.M.: 'Experience and progress
with UHF diagnostics in GIS. CIGRE, 1992, Paper 15/23-03
65
67