Figure 3.14 shows a selection of typical SF6breakdown characteristics for the elec-trode systems illustrated in Table 3.4 [14]. From the vast amount of breakdown data Generally it might be expected that at the
lower end of the voltage/current scale, where only one or two series breaks per phase are required, i.e. up to 420 kV the live-tank construction is probably more economical;
but at higher voltages with three or four
(or more) series breaks, the dead-tank con-struction tends to have an economic advan-tage because of the reduced amount of external insulation. It is worthwhile noting that dead-tank circuit-breakers are inherently more suitable for areas subject to earthquakes.
Figure E Live- and dead-tank circuit-breakers (Courtesy NEI Reyrolle Ltd) Table 3.3 (Continued)
Applications of gaseous insulants to switchgear 119
Typical duplicate bus circuit
Circuit-breaker 1
2 3 4 5 6 7 8 9 10 11 12 13
Interruptor Hydraulic mechanism Disconnecter (circuit) Disconnecter (bus selector) MES
FMES Gas console Current transformers Bursting disk Voltage transformer Reserve busbar Main busbar
Figure 3.12 Sectional view through a typical 420/550 kV GIS duplicate busbar circuit (Courtesy VA Tech Reyrolle)
Figure 3.13 Example of 420 kV GIS equipment on a recent substation project in the Middle East (Courtesy VA Tech Reyrolle)
Table 3.4 Electrode systems [14]
Electrode system
Description Shape Size (mm) HV electrode
material
*Various resin formulations have been studied, including, silica, alumina, bauxite and dolomite filled systems.
†Using corrugated outer cylinder.
Applications of gaseous insulants to switchgear 121
accrued, the limits of lightning and switching impulse withstand gradient perfor-mance for a family of large practical GIS gas-gap type configurations were produced and are illustrated in Figure 3.15 for SF6pressures in the range 0.1< p < 0.6 MPa.
Experimental determination of critical breakdown (E50) and highest withstand (Ew) gradient values was obtained using similar test techniques to those adopted pre-viously [14] (i.e. E¼ V/hg; see section 3.5.2). It should be emphasised that these curves present typical design-type gradient relationships and encompass the results obtained for a large family of coaxial cylinder and perturbed electrode configurations for gas-gaps in the range 50< g < 180 mm. Theoretical breakdown gradient levels are also given in Figure 3.15(a–d); it is readily apparent that practical results deviate from theory as SF6gas pressure increases [14]. A corresponding withstand curve for 50 Hz conditions is also shown in both Figures 3.14 and 3.15.
2.0 11 50 Hz (60 s withstand)
MV
Figure 3.14 Example of SF6breakdown characteristics [14]
(a)40 (b)
Figure 3.15 Typical limits of 50% breakdown gradient (E50) and critical withstand gradient (Ew) on SF6pressure for large coaxial and perturbed cylindrical electrode systems under clean condition [14]
Curves (a–d): lightning impulse (LI) waveshape (1.2/50ms) Curves (e–h): switching impulse (SI) waveshape (250/2,500ms) (- - - -) shown in curve (c): the lower limiting 50 Hz withstand
characteristic (Ew)
E50data: curves a, b, e and f; Ewdata: curves c, d, g and h
Applications of gaseous insulants to switchgear 123
To provide a better understanding of statistical variations possible with 50 Hz dielectric breakdown characteristics, relating to practical GIS-type assemblies, it is relevant to consider the following results [14]. Figures 3.16–3.18 show the variation in individual short-term 50 Hz breakdown performance, with repeated sparking, for large perturbed and unperturbed cylindrical electrode systems. For comparison, the corresponding 60 s highest withstand levels (Vw) are also given.
Some salient findings emerge.
1. Figure 3.16, curves (a–c), relates to a 50 Hz voltage test sequence for three
‘successive’ test series for a concentric cylindrical electrode system.
i. Curve (a) illustrates that whereas the highest ‘60 s withstand level’ Vwwas 480.8 kVpk, the sequence of ‘25 instantaneous’ breakdown levels was significantly higher, being within the range 520–595 kVpk. For the pur-poses of comparison, the theoretical critical breakdown voltage level (VTH) is 621.6 kVpk, based on a critical (E/p)lim¼ 89 kV mm1MPa1.
ii. Curves (b) and (c) show corresponding results for two repeat series. Here, it can be seen that the corresponding ‘60 s withstand’ levels (Vw) increased significantly to 565.7 and 594 kVpk, respectively, as compared to the level of 480.8 kVpkobtainable in the first test series, i.e. curve (a).
iii. Similarly, it can be seen from the sequence of individual ‘instantaneous break-down’ levels (Vb) that, despite occasional low-level breakdowns, a significant
‘conditioning’ effect has taken place and, in curve (c), approximately 10 of the 25 ‘instantaneous breakdown’ levels have achieved breakdown values of
600 kVpk, i.e. within 3.5% of the theoretical limiting breakdown level.
0
Figure 3.16 Sequence of 50 Hz breakdown levels in SF6for concentric cylinder electrode system F (SF6pressure: 0.1 MPa)
- - - VTHTheoretical breakdown level (621.6 kV) (Based on (E/p)lim ¼ 89 kV mm1MPa1)
~ Individual spark breakdown
*p ‘PIP’ – partial (incomplete) breakdown
Vw Maximum (1 min) withstand level, established immediately prior to test runs a, b or c, respectively
2. Figure 3.17 shows comparable 50 Hz results for a perturbed concentric cylin-der electrode system (F). As before, the results presented in curves (a)–(c), respectively, relate to three complete test series. Once again, the sequence of
‘instantaneous breakdown’ results exhibits a noticeable voltage ‘conditioning’
effect, with occasional low-level results, e.g. curve (c), the conditioned level of
590 kVpk being within 3% of the theoretical breakdown value. Returning briefly to the sequence of highest ‘60 s withstand’ levels for the test series (a)–
(c) (Figure 3.17), it is noted that Vwincreases as the test series proceeds, cor-responding values being 509.1, 544.5 and 558.6 kVpk, respectively.
3. Figure 3.18 provides comparable results, obtained at higher SF6gas pressures.
Under these conditions, ‘instantaneous breakdown’ values were significantly lower than corresponding theoretical levels, e.g. being lower at 0.2 MPa, by between 15% and 31%, for this electrode arrangement.
4. Figure 3.19 shows some further results for a larger concentric cylinder elec-trode system (I) (see Table 3.4), which illustrate similar trends at 0.1 MPa but show a very significant spread in the SF6‘instantaneous breakdown’ levels as compared to the ‘60 s withstand’ level (Vw) of 795 kVRMS, at the higher gas pressure 0.45 MPa (see curves (a), (d) and (e)). Here, the values of Vwtend to be lower than the theoretical limiting levels VTHby between 18% and 60%, the difference increasing with SF6pressure. It should also be noted (curve (a)) that the lowest recorded breakdown level corresponds to a stress figure of
8 kVRMS/mm, significantly higher than the working stresses for GIS equip-ment referred to earlier.
Figure 3.17 Sequence of 50 Hz breakdown levels in SF6for perturbed electrode system F (SF6pressure: 0.1 MPa)
- - - VTHTheoretical breakdown level (574 kV) (Based on (E/p)lim ¼ 89 kV mm1MPa1)
~ Individual spark breakdown
*p ‘PIP’ – partial (incomplete) breakdown
Vw Maximum (1 min) withstand level, established immediately prior to test runs a, b or c, respectively
Applications of gaseous insulants to switchgear 125
Extensive data exists for SF6gas-gaps under clean and contaminated condi-tions (see References 14, 25–30, for example). It is now established that the presence of particulate contamination of lengths 2–20 mm can reduce the dielectric withstand capabilities of practical gaps by varying amounts up to typically 30%, 40% and 70% for lightning impulse, switching impulse and power frequency conditions, respectively, at working SF6pressure. Figure 3.20 illus-trates a typical spread in 50 Hz flashover levels for varying degrees of gross contamination and represents the maximum lowering of withstand that can be expected [14].
Barrier performance data: Careful design and assembly of the cast resin sup-port barriers used in SF6-insulated GIS equipment are vitally important. It should be noted that, for a particular gas pressure, the withstand characteristics of support barriers in SF6under clean conditions depend on the particular resin formulation used, the insulation shape and the disposition of stress-relieving fitting, insert, etc.
Typical withstand gradient levels of 11.6, 8.7 and 6.6 kVpk/mm can be achieved under lightning, switching impulse and 50 Hz short-term voltage conditions, respectively [14].
0 700 800 900 800 900
1000 1000
kV
800 900 1000
5 VW = 869.7 kV VW = 862.7 kV VW = 799 kV
p
p p
p
b
c a
10 15 20 25
Shot number
Figure 3.18 Sequence of 50 Hz breakdown levels in SF6for perturbed electrode system F (SF6pressure: 0.2 MPa)
VTHTheoretical breakdown level (1,148 kV) (Based on (E/p)lim ¼ 89 kV mm1MPa1)
~ Individual spark breakdown
*p ‘PIP’ – partial (incomplete) breakdown
Vw Maximum (1 min) withstand level, established immediately prior to test runs a, b or c, respectively
The presence of particulate contamination can reduce the 50 Hz withstand capability of cast-resin support barriers in SF6gas by varying amounts (e.g. up to
<30%) depending on particulate size and disposition. For the most onerous dis-positions of cast-resin support barriers, the percentage lowering of ‘withstand’
performance under impulse conditions tends to be much less than that experienced for 50 Hz test conditions, for comparable levels of contamination (see Figures 3.20 and 3.21, for example) at spacer gas-gap interfaces. The reader should compare these
800
800
800 700
700
700
400
400 300
300 900 1000
0 5 10 15 20 25
Shot number
(0.1 MPa) (0.1 MPa) (0.3 MPa) (0.3 MPa) (0.45 MPa)
a
b
c
d
e VW = 360 kV
VW = 380 kV VW = 745 kV VW = 705 kV VW = 795 kV kVRMS
Figure 3.19 Sequence of 50 Hz breakdown levels in SF6for large concentric cylinders (electrode system I, SF6pressure: 0.1–0.45 MPa) D Individual spark breakdown
Vw Maximum (1 min) withstand level, established immediately prior to test run
Applications of gaseous insulants to switchgear 127
findings for 100% SF6with Figure 3.11, diagrams (e)–(g), relating to particle-initiated breakdown in SF6/N2mixtures. These workers [18] consider that particle-initiated breakdown is more complicated in SF6/N2mixed gas than in 100% SF6gas and they recommend that further studies are necessary to achieve a better understanding of the discharge mechanisms.
0 2 4 6
Clean withstand (1 min)
630 kV works test level 520 kV routine test level
420 kV system SF6 0.55 MPa
Phase voltage
420 kV site test level
8 10 12
Particle size (mm) Typical particle
lift-off voltage 200
400 kVRMS
600 800 1000
Figure 3.20 Fifty Hertz flashover characteristics of epoxy resin conical spacers under varying degrees of metallic contamination [10]
2.0
1.5
1.0
Lightning impulse (MVpeak) 0.5
0
Silica filler
Alumina filler
Barrier orientation Clean
Clean
Contaminated
Typical spread for reduction in Li withstand
Contamination
Metallic particles < 5 mm Non-conducting fibres < 10 mm
Figure 3.21 Limit of lightning impulse withstand capabilities of epoxy resin conical spacers under clean and contaminated conditions