freedom to make any measurements judged to be helpful, instrumentation; and A. Jeremenko, for operation of the cir-advances in the knowledge of circuit interruption can be cuit.made.

Conclusion References

[1] H. Edels and F. Fenlon, "A theory of interruption anid thermal-An experimental method has been developed which assists arc reignition," Proc. IEE, vol. 110, November 1963.

in advancing the understanding of are interruption phenomena [2] J. D. Cobine, E. E. Burger, and G. A. Farrall, "Impulse systemand the development of arc interrupters. This approach has for arc-recovery-strength measurements," Trans. AIEE (Power

Apparatus and Systems), vol. 77, pp. 4094-1100, December 1958.provided valid information with a minimum of cost. It [3] T. H. Lee, Allan Greenwood, D. W. Crouch, and C. H. Titus,derives its usefulness fromn the flexibility and accuracy of cir- "Development of power vacuum interrupters," Trans. AIEE

cuit c t t e d(Power Apparatus and Systems), vol. 81, pp. 629-639, 1962CUit conditions, the speed and ease of operation, the quality Of (February 1963 Section).instrumentation, and the reduction of cost considerations [4] J. Biermanns, "The Weil circuit for testing of high-voltage circuitfrom fundamental studies. breakers with very high rupturing capacities," Paper 102,CIGRE, Paris, France, 1954.

[5] W. F. Skeats, C. H. Titus, and W. R. Wilson, "Severe rates ofAcknowledgment rise of recovery voltage associated with transmission line shortThe development, design, and application of this circuit circuits," Trans. AIEE (Power Apparatus and Systems), vol. 76,

pp. 1256-1266, 1957 (February 1958 Section).has drawn on the talents of many people, and the authors [6] T. W. Liao, H. N. Schneider, W. F. Skeats, and C. H. Titus,wish to acknowledge the contributions of these individuals: "Switching of EHV circuits, IV-Compound circuit test facilityDr. T. H. Lee and CT. H. Titus, for initiating and encouragingfor high-voltage circuit breakers," IEEE Trans. on Power Appa-DJr. T . H. Lee and C. H. Titus, for initiating and encouraging ratus and Systems, vol. 83, pp. 1213-1222, December 1964.the work on this equipment; Dr. A. N. Greenwood, H. N. [7] T. Lee, A. Greenwood, and D. White, "Electrical breakdown ofSchneider, W. C. Kotheimer, and V. Mishkovsky, for contribu- high-temperature gases and its implication in post arc phenomenaY I,. ,, . . 1 in circuit breakers," presented at the 1965 IEEE Winter Powertions to the design of circuit components, controls, and Meeting, New York, N. Y.

Design and Application of EHVDisconnecting SwitchesArem Foti, Senior MemberIEEE

Abstract: Technical data, accumulated while designing and tion, it must provide the capacity to conduct all normal andtesting EHV (extra-high-voltage) disconnecting switches, are abnormal currents which flow in the electric system safely.being presented as an application guide in establishing EHV The disconnecting switch must have proper phase-to-phaseswitch insulation coordination. Dielectric characteristics oflarge single and large double switch gaps are included for making insulation, phase-to-ground insulation, and open gap insula-comparisons between single-break and double-break switches. tion. There also exists the requirement to operate theThe subject of EHV switch open gap insulation coordination is switch mechanically between open and closed positions in areviewed in rather complete detail. Design criteria and test positive dependable manner even under heavy ice accumula-data are submitted, covering a 500-kV switch employing a Y

variety of EHV insulators. Attention is called to several tions or under the influence of corrosive contaminants. Thesubjects requiring action and resolution by electrical-power- disconnecting switch for EHV stations is expected to performindustry standardizing organizations. all these functions and, simultaneously, require absolutely

minimum maintenance, repair, or inspection of its own com-Bulk electric power transmission has been expanded in recent ponents.years to include transportation of electricity at extra high Switch Construction Classificationsvoltage. Many transmission lines are now planned or in exist-ence at 500 kV in the United States of America and at 700 kV Several switeh types have been considered for the EHVacross the border in Canada. A necessary part of these range. The vertical-reach or pantograph-type shown in Fig.transmission systems is a switching station, best operation of 1 is applicable, but its use requires a special switching sta-which is dependent on selection and application of electric tion layout and electric bus arrangement. From this stand-whic isdepnden onselctin an aplictionof lecric point, it may be considered a special-purpose switch whichapparatus, including disconnecting switches used for isolating withit a relatively h prci on st.equipment or transmission lines at required times. carries with it a relatively high production cost.By providing a visual air gap, a disconnecting switch in the The horizontal-opening center-break switch of Fig. 2 would,

open position assures positive isolation of equipment and line by contrast, seem to represent lower production cost, since itsections, affording safe system examination, maintenance, is simple in design and has only two insulator columns per

and epar.hentheswich s eergzedin the closed posi- switch pole unit. But, associated with its use are heavy-dutyand repairWhentheswitch isenergizedinsulator support bearings, special terminal designs, and the

Paper 31 TP 65-75, recommnended and approved by the Switchgear need for greater phase-to-phase spacings, all of which tendCommittee of the IEEE Power Group for presentation at the to offset the apparent production econornies.February 5, 1965. Manuscript submitted November 2, 1964; made bigrea switch Frep useintaieotheEhVoraznge-io-pandn 1800l-available for printing November 30, 1964. beksic.Fruei h H ag-50 n 80AREM FOTI is with the I-T-E Circuit Breaker Company, Greensburg, kV BIL (basic insulation level)-some manufacturers arePat. promoting designs with three or four insulator columns per

IEEE TRANSACTIONS ON POWER APPARATUS AND SYSTEMS VOL. PAS-842 NO. 10 OCTOBER 1965

Fig. 3. D~ouble-<r lliilbreakhorizontal-opening switch

Fig. 1. EHV vertical-reach switch of type used in Europe

Fig. 2. Center-.break horizontal.opening switch Fig. 4. Single-break vertical-opening 500-kV switch used In

United States

pole. To the author's best knowledge, this switch construc- ages, with power frequency voltages, and with switching surgetion has not been considered for applications where voltages voltages. The single and doable gap test data are presentedexceed 1800-kV BIL. graphically in Fig. 5 and Fig. 6, respectively.The vertical-opening single-break switch type seen in Fig. 4 Examination of these data, in view of the phase-to-ground

has had the greatest use in practically any and all voltage and dielectric strength developed by the various EHV insulatorcurrent ratings in years past. Its prominence in the American columns in existence, resulted in a series of thought-provokingelectrical industry is of such magnitude that it needs no observations. Some of the most forceful were these:further introduction. It is planned for use with both pre- 1. In order to compare switch open gap dielectric propertiesviously mentioned 500-kV American and 700-kV Canadian with phase-to-ground insulation characteristics, there is needswitches. to standardize the method for conducting tests.Each switch construction classification described has its 2. By the same token, there is necessity to standardize the

advantages and disadvantages when one is compared with method for determining switching surge withstand voltages.another. However, since the latter two types have domi- 3. Agreemnent must be reached in the type, or form, of switchingnated the EHV scene in America and since these are the two surge wave to be used in laboratory tests.types that have engendered the most comment, a more de- 4. Rules must be established for uniform application of cor-tailed comparison is suggested. rection factors to translate switching surge test voltages to

standard conditions of relative air density (RAD) and humidity.Single Vertical Break Vs. Double Horizontal Break 5. The energy output of the impulse generator used for switch-

Essential for comparing the two switch forms named in the ing surge tests must be of sufficient magnitude to flash across

heading is more knowledge than heretofore obtainable con- the large gaps, or misleading results are obtained.

cerning the dielectric characteristics of single switch gaps and Brief comments on these five points will further clarifydouble gaps. In quest of this knowledge, a comprehensive the thoroughness with which comparisons were made. Alldielectric testing program was conducted on extra-large tests were conducted indoors, in the spacious laboratory ofswitch gaps when stressed with lightning impulse-type volt- Centro Elettrotecnico Sperimentale Italiano (CESI) at

1965 FoTn: EHV DISCONNECTING SWITCHES 869

2800 -_ __-__ __ __ __-__

2700- -- I

2500 --_ _

2400 < 0 _____

2300-

2200p___4__

21 00 T -__

2 SWITCo t LRGINEGATIV l , -- - g-----:DRY 90'FLASHOVER /-

O 0

DRY IO%FLASHOVERx> WET90%FLASHOVER i

J80 0 _s WET 10 %FLAStOVER ,x//+|{*!*0**_J 1800 W IFLAi OVERX SWITCHINGSURGEPOSITIVE-DRY

90% FLASHOVER

700 /-

< 1600--- -

SWITCHING SURGE POSITIVE- DRYo 'I110%FLASHOVER> I 50°° --t 4..,f,_

400 SWITCH1INGSURGE POSITIVE . -tt

WET 90%FLASHOVERX

30--oo -_--.--- ----t--t -4----- A_ __t- _60-C/S DRY FLASHOVER

WET 10% FLASHOVER X; MAXIMUM (RMS)

12004 __--_ _t670-C/SWETFLASHOVER MAXIMUM(RMS)

I0C/S DRY FLASHOVERI100 - _I_!___ MINIMUM(RMS)

._c6O-C/SWETFLASHOVER MINIMUM (RMS)

1000 _-_ __ 1 __ __ __ _ _ _ _ _ -__ _ _ i _ _ _ _900j _X F

ALL VOLTAGE VALUES ARE CORRECTEDTO STANDARD CONDITIONS

800 i T h

X-DENOTES ACTUAL TEST VALUES

80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230INCHES-OPE-N GAP (METAL TO METAL)

SINGLE BREAK SWITCH-VERTICAL OPENING

Fig. 5. Graph for single gap dielectric properties

870 IEEE TRANSACTIONS ON POWER APPARATUS AND SYSTEMS OCTOBERh

2800

2700-

2600 -

2500 i1

24001 t t_

NEGATIVE 1MPULSE90%FLASHOVER X

2 300 POSITIVE IMPULSE 90% FLASHOVER- x LPOSITIVE IMPULSE 10%FLASHOVER. x

_______NEGATIVE IMPULSE 10%FLASHOVER x _ _2200-

21 00

210I~~ 9 o 0I~J sos 11___ _ I4 __l

° SWITCHINGWSURGENEGATIVE-DRY90%4FLASHOVER x> SWITCHING SURGE NEGATIVE-WET 90%FLASHOVER xI0

J18 00l

SWITCHING SURGE NEGATIVE-DRY 10I FLASHOVER x

Z 700 I - _SWITCHING SURGE NEGATIVE-WET 10%FLASHOVER xC

1600 _ t

.___ SWITC IN SUI5001 . a i , | , l / POSITIVE -WET

10 %FLASHOVER

_____l I /SWlTCHING SURGE1400' 1I ~- POSITIVE-DRY

140-- IO10 FLASHOVERSWITCHING SURGE POSITIVE-DRY 90%FLASHOVER-

I30CaSWITCHING SURGE POSITIVE -WET 90% FLASHOVER- ,

1200--- 4 _ _ __ _

0,S , /

60-C/SDRYFLASHOVER MAXIMUM(RMS)j1000- 60-C/SWETFLASHOVER MAXIMUM(RMS)' x-' -_10O o- 6-C/ DRY FLASOE MNTUMRMI)

60-C/S DRY FLASHOVER MINIMUM(RMS) iw60-C/S WETFLASHOVER MINIMUM (RMS) x-

T

I | ALLV6LTAG VALUESARE CORRECTED__ __I__ __ ¶tO STA 'DARD COND IONS

800-X-DE10JOTES CTUAt TEST VALU S

700 t80 90 100 liO 120 130 140 150 160 170 180 190 200 210 220 230

INCHES-OPEN GAP (METAL TO METAL)TOTAL TWO GAPS IN SERIES-DOUBLE BREAK SWITCH

Fig. 6. Graph for double gap dielectric properties

1965 FOTI: EHV DiSCONNECTING SWITCHES 871

exists a pronounced difference, for example, in a switch witha nominal voltage rating of 500 kV. Table I shows the resultsof analyzing the two philosophies when considering a single-break switch of the Fig. 4 variety, assembled on 1800-kVBIL pin-and-cap-type insulator columns. When the insula-tion coordination is conducted by PHILOSOPHY A method, posi-tive-polarity impulse voltage is the controlling voltage stress,and the required open gap could be as small as 138 inches.With the PHILOSOPHY B approach, the dry positive-polarityswitching surge voltage stress is the determining factor, andthe open gap should be 216 inches.

This more than 6-foot difference in switch open gap dimen-sion has given rise to a new philosophy that merits considera-tion. The new proposal analyzes phase-to-ground flashoversand switch open gap flashovers on a probability basis. Itrequires that the voltage which will produce 90-percent prob-ability of flashover of the switch jaw or hinge insulator as-sembly will produce less than a 10-percent probability ofswitch open gap flashover.For example, Fig. 8 shows this basis of switching surge

open gap coordination. Observing that the voltage whichproduces 90-percent probability of flashover of the insulatorcolumn is 1450 kV, the same as recorded in Tables I and II,this same 1450 kV will produce a probability of flashover ofless than 10 percent for a 168-inch single gap. This basis ofswitching surge open gap coordination shows prospects ofbeing a good compromise in maintaining switch insulationcoordination, without using either excessively long switchblades or appendages to control insulator column dielectric

Fig. 7. Phase-to-phase flashover double-break switch during strengths.laboratory tests Table II represents a study of pertinent insulation charac-

teristics for a double-break switch when coordinating theswitch open gap with the insulation characteristics of 1800-kVBIL pin-and-cap insulators. The extremely large difference

Milan, Italy. Voltage values appearing in the graphs of in switch open gap dimensions derived from employment ofFigs. 5 and 6 were obtained when testing in accordance with the two philosophies under discussion is a possible explana-the recently proposed method for determining switching tion for the widely divergent views expressed by switch manu-surge withstand voltages [1]. facturers regarding the practicability of the double-breakA switching surge wave shape, commonly described as a switch for EHV voltages.

250/3250-Mjs wave, was used [2], [3]. Switching surge volt-age values were corrected to standard conditions for RADand humidity, being treated as though they were impulse Switch Design and Testsvoltages. The energy output of the impulse generator was After designing a single-break vertical-opening switch pro-sufficiently large to produce a switching surge flashover, as totype in physical dimensions, proportioned to the conceptshown in Fig. 7, when testing two phases of a double-break- that the voltage producing 90-percent probability of insulatortype switch assembled on 25-foot phase centers. Each column flashover would produce less than 10-percent prob-switch blade is extended 10 feet into the phase clearance, ability of open gap flashover, an extensive dielectric testingsimulating a 750 blade opening. program was consummated. The test data have been assimi-

lated and are represented in Figs. 9 through 11. In order notPhilosophy of Switch Insulation Coordination to be cumbersome, the data are tabulated for withstandRecalling present-day industry standards [4] for deter- voltages only.

mining switch length of break (or open gap, as it is commonly The need for standardizing a method of conducting tests wascalled), but not quoting the section verbatim, it is established evident. Subjects requiring special consideration relate tothat an outdoor air switch must be such that the open gap height of test object from ground floor, proximity of adjacentwill withstand various types of voltage stresses that are 10 ground planes or walls, shape of switching surge wave to bepercent in excess of a prescribed phase-to-ground with- used in laboratory tests, method of determining withstandstand voltage. For the sake of clarity, let us call this PHILOS- voltage levels, effect of output energy rating if the impulseOPHY A. generator on test results, and rules for applying correctionEven though PHILOSOPHY A is the accepted industry stand- factors to voltages for switching surge tests. A by-product of

ard, it is common knowledge that many switch users specify the switch testing program is the debate concerning relativethat the switch open gap must be of such size that it will merits of indoor and outdoor tests. Even though favor-withstand voltages that are 10 percent in excess of phase-to- able comments may be made for both, this paper merelyground flashover voltages. This latter philosophy of switch calls attention to situations which eventually will requireinsulation coordination we will call PHILOSOPHY B. resolution.

In subtransmission voltage ratings, employment of either In addition to dielectric tests, the 500-kV switch passedmethod does not present serious difficulty. However, there continuous current rating and momentary rating tests.

872 IEEE TRANSACTIONS ON POWER APPARATUS AND SYSTEMS OCTOBER

Table 1. 500-kV Single-Break Switch Insulation

Coordination According to PHILOSOPHY A (Present NEMA SG6 Standard)Selected Insulation

Selected Insulation Withstand Plus 10 Switch Open GapWithstand Voltage, Percent in kV (Open Distance Required,

Type Voltage Stress kV Gap Withstand) Inches (from Fig. 5) Remarks

60 c/s, dry 810, rms 891, rms <120~Switching surge positive, dry 1125 1238 <120 Positive impulse voltage controls. Gap couldPositive impulse 1800 1980 138(be as small as 138 inchesNegative impulse 1800 1980 1251

Coordination According to PHILOSOPHY BVoltage Value Previous Column Plus Switch Open Gap

90-Percent Flashover, 10 Percent (Required Distance Required,Type Voltage Stress 10-Unit Pin-Cap Open Gap Withstand) Inchest (from Fig. 5)

60 c/s, dry 1030, rms* 1133, rms*Switching surge positive, dry 1450 1595 216Positive impulse 2110 2321 162Negative impulse 2305 2535 164* Sustained 60-c/s voltages of this magnitude virtually impossible to obtain in service.t Determined by extrapolation.

IOC____ _ 8 KVBlLPINCAPINSULATOR

~9.

0 o.176"l DOUBI1,E GAPGo ~~~~~~~~200"QOUBLEG~AP _

>-50~~ ~ ~ ~ ~ ~ s - To

SWITCH ING SURGE -KV

Fig. 8. Graph for insulation coordination, dry positive-polarity switching surge, single- and double-break, 230/3250 As

Table II. 500-kV Double-Break Switch InsulationCoordination Accordirng to PHILOSOPHY A (Present NEMA 5G6 Standard)

Selected InsulationSelected Insulation Withstand Plus 10 Switch Open GapWithstand Voltage, Percent in kV (Open Distance Required,Type Voltage Stress kV Gap Withstand) Inches (from Fig. 6) Remarks

60 c/s, dry 810, rms 891, rms <152Switching surge positive, dry 1125 1238 158 Positive dry switching surge controls. TotalPositive impulse 1800 1980 <152(two gaps could be as small as 158 inchesNegoative impulse 1800 198 <152

DRY SWITCHING SURGE WET SWITCHING SURGETEST WITHSTA N D KV* WITHSTAND KV *

CONDITION IPOLRI YPIN HOLL SOLID POLARITY PIN HOLLOW SOLIDI|POLARITY |CCAP POST FPOST CAP POST POST

A S A POSITIVE 1277 1360t 1525 POSITIVE 1222 |lOt 1425IINSULATOR |NEGATIVE |b8I 2125t 2130 NEGATIVE 1354 1200t 1770

SWITCH POSITIVE 1170 1470 1440 POSITIVE 1350 1250 1475CLOSED

NEGATIVE 2040 1650 2220 NEGATIVE 1470 1280 1840

SWITCH OPEN POSITIVE 1230 1400 1360 POSITIVE 1250 1330 1475JAW TEST

NEGATIVE 2150 - 2220 NEGATIVE 1470 1200 1840

SWITCH OPEN POSITIVE 1320 1350 1290 POSITIVE 1250 1220 1320HINGE TEST

NEGATIVE 2140 1840 2330 NEGATIVE 1420 1200 1800

RATED I-T -E 2 S7 a 112SKV I-T-C 2! - 1125 KVVALUE POSITIVEORNEGATIVE WAVE POSITIVEOR NEGATIVE WAVEI

Fig. 9. Electrical test data: 500-kV switch and bus insulators

Switching surge dielectric test data for 500-kV switches and bus supports. Withstand values are for two sigma prob3bility or better.Switching surge tests conducted with 250/4000-,us wave. Wet tests conducted in accordance with ASA Standard C 29.1. Base of insula-tor column 12 feet from ground plane for all tests. Cap and pin insulators consist of five TR 58-9 units over five TR 58-9 alumina units.Hollow post insulators are 5-unit tapered assemblies. Solid post insulators are 3-unit assemblies. All tabulated voltages values havebeen corrected to standard conditions. Values markedt obtained from published information; the test basis of these values is notknown

IMPULSE WITHSTAND 50-C/S WITHSTANDTEST KV\* KV (RMS)ft

CON DITIONPIN HOLLOW SOLID TEST PIN HOLLOW SOLIDPOLARITY CAP POST POST CONDITION CAP POST POST

A S A POSITIVE 202 1 2025t 2025 DRY 925 1ooot 855jNSU LATOR

NEGATIVE 2171 2225t 2470 WET 758 790t 690

SWITCH POSITIVE 1960 2180 2120 DRY 930 940 950CLOSED

NEGATIVE 2060 2330 2340 WET 800 760 840

SWITCH OPEN POSITIVE 1920 2240 2020 DRY 900 975 990JAW TEST NEGATIVE 1970 2320 2340 WET 810 87 0 915

SWITCH OPEN POS1TIVE 1980 2180 2020 DRY 905 930 855HINGE TEST

NEGATIVE 20860 2250 2320 WET 780 800 870RATED 1800KV 810 KV DRYVALUE POSITIVE OR NEGATIVE WAVE 7 1 0 KV WET

Fig. 10. Electrical test data: 500-kV switch and bus insulators

Impulse and power frequency test data for 500-kV switches and bus supports. Withstand values are for two sigma probability or bet-ter. Impulse tests conducted with nominal 11/2x40-pus wave. Wet tests conducted in accordance with ASA Standard C 29.1. Base ofinsulator column 12 feet from ground plane for all tests. Pin and cap insulators consist of five TR 58-9 units over five TR 58-9 aluminaunits. Hollow post insulators are 5-unittapered assemblies. Solid post insulators are 3-unit assemblies. All tabulated voltage valueshave been corrected to standard conditions. Values marked t obtained from published information; the test basis of these values isnot known

874 IEEE TRANSACTIONS ON POWER APPARATUS AND SYSTEMS OCTOBER

ConclusionRADIO INFLUENCE VOLTAGE

TEST MICROVOLTS EHV disconnecting switches are designed to meet bothCO N D IT I O N S K V TO P I N HOLLOW SO LID electrical and mechanical application requirements. Design

GROUND CAP POST POsr and application are virtuallv inseparable. A 500-kV dis-

A S A 3 2 O - 14 O 7 8 connecting switch of the single-break vertical-opening typeIN S ULA T O R has been designed, built, and tested, and pertinent test data

340 - 230 122 are submitted.32 0 1 52The single-break switch has been selected for EHV service,

CLO SED 320 lI 8 52 18 500 kV and higher, because it offers excellent switch open gap340 138 97 25 insulation coordination without requiring means for control-

ling the insulator dielectric strength.SWITCH OPEN 320 438 48 78 Single gap and double gap electrical characteristics for

340 560 205 I22 large EHV gaps have been presented as a contribution toimproving the art and increasing knowledge of air insulation

SW IT CH OPEN 32 0 28 0 36 2 7 in extra high voltages. These data should prove significantlyHINGE TEST 340 415 5l 153 helpful as a guide in determining switch open gap insulation

coordination.RATED AT 340 KV TO GROUND A new philosophy has been suggested to establish switchVA LU E MAX. 500 MI C ROVO LTS open gap insulation coodination. This philosophy provides

that voltages having a 90-percent probability of creating line-Fig. 11. RIV dielectric test data 500-kV switch and bus supports to-ground flashovers will, at the same time, have less than a

Base of insulator column 12 feet from ground plane for all tests. 10-percent probability of producing switch open gap flash-RIV tests conducted in accordance with NEMA 107. Insulators: overs. A thorough investigation of this subject by thePin-cap consists of (5) TR.58-9 units over (5) TR.58-9 alumina; electrical power industry is recommended for the purpose ofhollow post of (5) unit tapered assembly; solid post is (3) unit establishing a standardized way of determining insulationassembly. RIV values are averages of 4 tests each setting coordination.

There exists a need for establishing realistic and acceptableRI (radio influence) limits for EHV electrical apparatus. Itis encouraging to report that action on this subject has beeninitiated recently by the Technical Committee of the PowerSwitching Equipment Group of the National ElectricalManufacturers Association (NEMA).While conducting tests on EHV equipment and while

trying to draw comparisons between similar tests, it was con-cluded that much remains to be done by electrical industrystandardizing organizations. Included in the decisions to bemade are.

1. Conditions for conducting EHV switch and bus supporttests, such as height of test object from ground floor and mini-mum distance from other planes, and rules for applying cor-rection factors to switching surge voltage test values regardingrelative air density and humidity.2. Method for determining switching surge withstand voltage.3. Form of switching surge wave to be used in laboratory test-ing as related to rise time to crest voltage and decay time to

Fig. 12. Photograph taken during ice test of 500-kV single- one-half crest voltage.break switch

Significant though the progress has been in accumulatingknowledge and technical data in the EHV field, it appearsthat we have merely started; that much more work lies ahead

Figure 12 reproduces a photograph made during ice tests and much more knowledge must be obtained.which ascertained the ability of the switch to operate success-fully when coated with 3/4-inch thickness of ice. Theoperating mechanism, described in a previously presented Refereneespaper [5], was also used to operate the switch during switch- voltage[ for EHVinTuldtiation s osftmswitchIg surge withstanding surge tests at 500 kV [6]. The mechanism employs Apparatus and Systems, vol. 83, pp. 263-266, March 1964.torsional interphase rods, permitting rotation of the drive [21 J. W. Kalb, "How the switching surge family affects line insula-

tion," IEEE Trans. on Power Apparatus and Systems, vol. 82,insulator column through 180 mechanical degrees for ease of pp. 1024-1031, December 1963.operation and harmonic blade dampening when approach- !31 E. W. Boehne and G. Carrara, "Switching surge insulationing either open or closed positions-all highly desirable fea- strength of EHV line and station insulation structures," Paper

tures of EHYdisconncting switches. [4]415, CIGRE, Paris, France, 1964.tures of EHV disconnecting switches. [41 NEMA power switching equipment standards, Standard SG6,This vertical-break switch is ideally suited for adding tran- National Electrical Manufacturers Association, 1960.

sient voltage-limiting resistors [6], in case resistors become [5] Arem Foti and E. A. Williams, "Switches for 300 to 500 kV-some new design concepts," Trans. AIEE (Power Apparatus

necessary for controlling switching surge voltages. and Systems), vol. 73, pp. 1237-1247, October 1954.

1965 FOTI: EHV DISCONNECTING SWITCHES 875

[61 Arem Foti and J. M. Lakas, "EHV switch tests and switching S. C. Killian (Lapp Iinsulator Company, LeRoy, N. Y.): Thissurges," IEEE Trans. on Power Apparatucs and Systems, vol. 83, paper makes a good contribution to the literature in presentinigpp. 266-271, Mlarch 1964. curves on the electrical strength of long air gaps, both double

and single, and again pointing out the many conditions andmethods needing standardization in switching surge techniques.This standardization cannot arrive too soon; the present con-fused methods of writing specifications bear witness to this.

Since the switching surge wave is not as well behaved as theDiscussion familiar 60-cycle or impulse waves, some leeway in specifying

switching surge withstand values must be agreed upon. Theone, two, and three deviations on a Gausian distribution methodhave definite merit. On the other hand, a more readily under-stood method is that where a given piece of equipment should

E. A. Williams (MIemco Eniginieerinig anid Manufacturing Com- withstand 90, 95, and 98 percent, or whatever the user maypany, Inc., Commack, N. Y.): I wish to support very strongly decide upon, of a given number of shots at the voltage valueMr. Foti's proposed philosophy of determining the open gap desired.distances for disconnecting switches. His probability basis of Mr. Foti's suggested PHILOSOPHY B is similar to the lattergap determination is practical, is reasonably economical, and method in specifying that a voltage which will flash the insulatorshould provide adequate safety in switch gaps. In contrast, the 90 per cent of the time will flash the gap only 10 percent of thestandard method that he calls PHILOSOPHY A offers the user time. This seems a sensible philosophy for personnel protection,nio assurance whatever that the switch gap has a higher flashover and it should receive careful consideration for adoption by thethan the supporting insulators. Also, the method which Mr. committees now working on this problem.Foti calls PHILOSOPHY B may be totally impractical and un- Papers such as Mr. Foti's are invaluable in providing additionaleconomical as it basically requires a much larger and more information on station insulation and focusing attention on areasexpensive switch than either of the other two methods of gap which need standardizing at an early date.determination.

I recommended that the subject of switch gap determinationbe placed on the agenda of the appropriate NEMA and IEEE Joseph M. DeSalvo (West Penn Power Company, Greensburg,committees for consideration of early revisions in existing Pa.): Open gap coordination, referred to as PHILOSOPHY A, isstandards. correct but also misleading. Although NEMA SG6, 1960, does

I also support Mr. Foti's plea for standardization of EHV test specify that the open gap will withstand voltages 10 percent inmethods and procedures. In the early days of impulse testing, excess of the switch phase-to-ground withstand voltage, this isan interlaboratory coordinating committee arranged to have not the limiting criteria for the switch open gap design andseveral laboratories run tests on the same specimens; the results coordination. The standards also specify minimum open gapwere analyzed by the committee and the differences reconciled. metal-to-metal dimensions that, when used, provide for 20- toWe need this same approach for EHV. 30-percent margins. Reference can be made to paper [1] by

Noting that the switching surge tests reported by Mr. Foti G. E. Hertig wherein graphs are presented that contain with-were made with a 250/3250 wave in the case of Fig. 8 and with stands and flashovers of NEMA insulators and NEMA open gaps.a 250/4000 wave in the cases of Figs. 9 and 10, I wish to ask, The paper mentioned four switch types that have been con-"Why the change?" and, "Were the results significantly different sidered for use in the EHV range. The classification could bewith the two waves?" expanded to five types if the vertical-opening double-breakOur company recently had occasion to test some 500-kV bus switch were added to the list. This type has no rotating in-

supports. One object of the tests was to compare the per- sulator stacks and its operation is by hydraulic control.formance of the bus supports with the performance of switches Tables I and II show coordination of 500-kV switches, usingto be installed in the same switchyard. The switches had been PHILOSOPHY A and PHILOSOPHY B. Under PHILOSOPHY A, thetested previously in another laboratory. Much to our surprise, withstand voltages are listed that appear to be withstands inwe found that the two laboratories used different test procedures the order of 2.3-pereent probability of flashover. Withstands ofand different wave shapes. We still do not know whether or not present industry standards are in the order of 0- to 16-pereentthe test results were comparable. probability of flashover because one can have 0 for 3 or 1 in 6With regard to the RIV data given in Fig. 11 of the paper, flashovers for acceptability. If a higher probability of flashover

these seem to indicate a substantial difference in the RIV per- (such as 16 percent) had been used, a higher open gap distanceformance of cap and pin; and hollow- and solid-post insulators. would be required than has been shown.Our experience indicates that good, bad, and indifferent insu- The tables use pin-and-cap insulator 90-percent flashoverlators can be and are produced in any design. One of our bus values under the PHILOSOPHY B analysis. It would be worth-supports with cap-and-pin insulator produced an RIV of only while to mention what could be expected if hollow-post or solid-20 uV when tested at 380 kV to ground. We know of other post-type insulator 90-percent flashover values had been used.tests in which similar bus supports produced RIV of several Would this necessitate greater open gap distances for the samethousand ,uV. Be that as it may, the RIV of any 500-kV bus type of coordination because they generally possess highersupport or switch is a function of both insulators and fittings. flashover values than the pin-and-cap insulators?There is no standard or even accepted method of testing one As can readily be seen, if no control is used to limit maximumwithout the other. This subject also needs consideration by flashovers occurring across the phase-to-ground insulation, thenthe appropriate industry committees. the horizontal-opening double-break switch must have a muchMr. Foti shows four types of switch design in his Figs. 1 larger open gap than a vertical-opening single-break switch.

through 4. He dismisses lightly the vertical-reach switch and At 1800-kV BIL levels, the manufacturers of horizontal-openingthe center-break switch and proceeds to show the vertical-break double-break switches and possibly vertical-opening single-switch to be electrically far superior; he also comments that break switches will have to rely on the installation of effectivethese two designs have dominated the EHV scene in America. ground planes or rod gaps to reduce the maximum flashovers ofMy reaction is, "A plague on both of their houses!" For 500 the phase-to-ground insulation, if it is required that they providekV, the double-break switch leaves mueh to be desired elec- open gap coordination of the type recommended by the authortrically, and the vertical-break switch presents many mechanical (i.e., 90 percent flashover of insulation = or <10 percent flashoverproblems. Mr. Foti recognized this at least 10 years ago when of gap).he and I co-authored a paper, cited in this paper as [5]. Although the open gap comparison made in this paper is note-The vertical-reach switch (Fig. 1 of the paper) and the center- worthy, the most significant difference or comparison between the

break switch (Fig. 2) have been widely applied for a number of two is the respective required phase-to-phase spacings of the twoyears in Europe at voltages of 460 kV and higher. Both of these types of switches. I feel that this is a significant factor for com-switches are electrically better than the double-break switch parison between the two types.and mechanically better than the vertical-break switch. The The open gap philosophy suggested in the paper as a recoin-potential purchasers of EHV switches will do well to make a mended procedure in EHV applications appears adequate ifcareful engineering analysis of the advantages and disadvantages the voltage values being compared can be considered accurateof the various switches available to them, to a high degree of confidence. Present data available from

876 IEEE TRANSA~CTIONS ON POWER APPARATUS AND SYSTEMS OCTOBER

EHV switch tests show inconsistencies in laboratory techniques Chairman of the Air Switch Working Group of IEEE, to reportand in philosophies of preparing the data. Reference may be on the activities of the Working Group in the matter of switchingmade to a paper [2] wherein a fresh approach is taken that surges. Using as its basis for starting to work, a paper [21 onprovides for a rigorous and controlled test employing probability switching surge insulation determination, presented by Mr. J. B.determination and using the normal (Gaussian) distribution Owens at last year's meeting, the Working Group has arrivedcurves. The method suggests that voltage values be stated in at a proposed standard for switching surge testing of air switches.the form - kV4- kV (error possibility) with first-, second-, This document has been approved by the Switchgear Committeeor third-order confidence range. I suggest that the coordination for letter ballot on the basis of a 1-year trial period. Some ofphilosophy recommended by Mr. Foti be expanded such that the more pertinent points in the proposed standard include:the open gap coordination be stated as follows: "The phase-to- stipulation of height of the test piece above ground, proximityground insulation 90-percent probability of flashover voltage of other ground planes, number and magnitude of test voltage(- kV+- kV with X order confidence) be less or equal to the applications, definition of wave shape, method of determiniingopen gap 10-percent probability of flashover voltage (- kV- insulation characteristics by use of probability paper, correction- kV with X order confidence). factors, open gap coordination, and mathematical method of

evaluating test data.REFERENCES The ballot will be in the mail in February. The Working[11 G. E. Hertig, "Disconnect switch insulation and coordination," Group solicits approval of this proposal on the basis that theTP-1240-3. industry needs a standard to provide a uniform approach to[2] G. E. Hertig and R. W. Ingman, "Fundamentals for the de- switching surge testing. Wide variations have been noted intermination of EHV switching-surge ratings," IEEE Trans. on data gathered thus far. Even though some detail disagreementPower Apparatus and Systems, vol. PAS-84, pp. 236-243, March may be found with portions of the proposed standard, if it is1965.

approved and used for a 1-year trial period, we will have gatheredsufficient data to either establish the validity of the standard orC. W. Upton (Westinghouse Electric Corporation, East Pitts- have a basis for its correction. The important thing is to have

burgh, Pa.): The electric utility industry's pursuit of 500-kV uniformity of environment and test procedure in the very ex-systems for economical transmission of power has posed many tensive testing that will be done in the next several years of EHVnew problems for the utilities themselves and for the apparatus development. This, we believe, will provide a real service tomanufacturers. Mr. Foti's presentation oIn the background, the industry.design philosophy, testing, and application of 500-kV switchesis to be commended as an accurate portrayal of what has been 5. Finally, wearing my Westinghouse hat once again, I wish toaccomplished thus far in solving this new engineering challenge compare in a broad sense the data which we have gathered withas well as in providing food for thought on work yet to be done. the data presented by Mr. Foti. Referring specifically to Fig. 5

I wish to present five points for consideration: in Mr. Foti's paper, I find that our impulse and 60-cycle dataagree well. Our switching surge data agree as far as slope is

1. I was most interested in reading Mr. Foti's analysis of various concerned, when plotted on probability paper, but Westinghousetypes of switches that might be considered for 500 and 700 kV levels are generally about 10 percent lower. Our testing wasand in following his thoughts on the merits or disadvantages of done in a smaller indoor laboratory than the CESI laboratory,each type. Recalling our deliberations, when we embarked on which may be considered comparable to outdoor testing becauseour development of 500-kV switches at Westinghouse a few of its large size. As Mr. Foti has indicated, there are validyears ago, we considered the same types of switches and came to arguments for testing on either basis and, at this stage of thethe same conclusions as Mr. Foti. We had considered one other art, no one knows for certain what the true answer really is.type; specifically, the double-vertical-break switch, which was I present this point merely to emphasize the desirability ofexcellent from the viewpoint of performance and reliability. getting into print a standard which provides guidance to allHowever, this design was judged to be a more feasible approach for a uniform approach to test parameters which have a signifi-to higher voltages than the 500-kV level, where conventional cant influence on the results obtained. In addition to thedesigns appeared satisfactory. inmportant point of proximity of the ground plane, such factors2. As to determining switching surge characteristics, our belief as wave shape, rate and resistivity of water, correction factors,based on extensive switching surge test programs is that switch- and definition of withstand, all influence the answers obtained.ing surge performance can be evaluated properly only after Approval of the proposed standard, discussed in my precedingreviewing large amounts of data. Our experience, based on a point, will help greatly to eliminate some of the variation nowlong test program, establishes the switching surge characteristic present in data gathered in different testing laboratories.slope of a plot on probability paper as being a 5-percent changein voltage per standard deviation. The slope as determined by REFERENCESshort isolated test runs, on any given day, may range from 2 [1] J. J. La. Forest, "The effect of station radio-noise sources onipercent per sigma to 7.5 percent per sigma; hence, our con- transmission line noise levels," IEEE Trans. on Power Appja-tention that a large test sample needs to be reviewed to establish ratus and Systems, vol. PAS-84, pp. 833-838, September 1965.an understanding of the true slope and position of the probability [21 J. B. Owens, "The determination of switching surge withstandplot in order to properly evaluate switching surge insulation voltages for EHV insulation systems," IEEE Trans. on Powercharacteristics. Even a 40-test-per-point probability plot may, Apparatus and Systems, vol. 83, pp. 263-266, March 1964.on occasion, deviate significantly from the 5-percenit slopeestablished by several thousand tests. R. M. Milton (Tennessee Valley Authority, Chattanooga,3. Mr. Foti urges the establishment of realistic and acceptable Tenn.): Mr. Foti is to be comnmended for outlining the problemslimits on RI. I would support him on this point and would urge facing the industry and the standardizing bodies on EHV dis-mts ondustr Iotwould support himsefon t uisintgandnwouldsurge connecting switches. The advantages and disadvantages ofthe industry not to penalize itself by requiring unnecessarily various types of switches are given, as well as the lack of testlow levels of RI for switches. On switches of this size, just the values and prescribed methods for performing design tests.ordinary practices of mechanical -design, adherence to good Although the vertical-reach or pantograph-type of disconniectfoundry practice, etc., provide an inherent RI level of probably switch was included, it was passed over rather lightly without1500 IA. Since there is no organic insulation associated with the serious consideration it seems to deserve. Many substa-switches, there is no deterioration whatsoever at this RIlevel. tion arrangements and layout.s were examined and studied by

,Von all three phases will not even be detected as radiated noise TV ansueqntydcredbasehyfildopoveimmediately outside the station. I would urge the industry Ase thmoes ftudescontinued,i becamues aprn.htotohto retain the levels which are presently in the standards at lower Astetuiscnneditbamaprnthtmotftevoltages. Extremely low levels of RI can be obtained with problems involved in securing a layout incorporating these

attndat icresen csts bu teseareeoss wichare inmy features would be solved if the movement of the disconnectattendant innceaeesrindcss bnutsthfeseaecstdhc.aenm switch blades were upward to close and downward to open.unjustified.Such an arrangement would leave the lower level of buses com-4. Mr. Foti properly stresses the need for establishing uniform pletely dead if the disconnect switches were in the open positionl.standards throughout the industry for testing in this new area This arrangement would allow vehicles or other mechanicalof switching surges. I would like to take this opportunity, as equipment to be moved inlto the deenergized portion of the

196 FOTI: EHV DISCONNEC rINC SWITCHES 8S77

substation in order to provide mechanical help in the mainitenance applicable to the vertical-break switch since both are of theof breakers or o)ther equipmenit. single-break type.The blade of the pantograph-type switch does move downward Again, in each case, mechanical ice-breaking capabilities haveto open and upward to close; consequently, it met our require- been proven with ice as thick as 21/2 inches in some cases andmenits and was selected for use in all 500-kV stepdown substations with the use of hydraulics which not only affords extremelybeing constructed by TV'A. In addition, it was found that the good control but also maximum output at minimum effort ifwidth of the substationi yard could be reduced about 100 feet manually operated. The use of hydraulic control for EHVby using vertical-reach switches instead of the conventional switches is also ideal because more than one switch may bevertical-break switches, anid this should reduce costs substantially. operated, either manually or remotely through motor-operatedThe author states that the open gap of a 500-kY disconnect pumps and at greater economic levels than previously availableswitch, designed in accordance with PHILOSOPHY A, could be with normal mechanical devices.

as smiall as 138 inches; if designed onl the basis of PHILOSOPHY B,the corresponiding dimenision would be 216 inches. It would beinterestinig to know whether the tests upon which Mr. Foti'sconclusions were based were made on extrapolated models of T. A. Burdeshaw (Southern States, Inc., Hampton, Ga.): Pub-lower voltage switches with small diameter blades and contacts lication of the test data presented in this paper constitutes aequipped with coronia shields. Some tests indicate that there substantial addition to the store of basic engineering knowledgeis good rewsoni to believe that the open gap dimension can be being developed. The cost of testing such as described in thissubstantially reduced by properly designed energized parts paper is extremely high, and the industry at large must sharewhiche providle genlerous contours and diameters.Iwaminsrovidenueieous oppositiono t aeauthor's. its findings to assure the economic application of EHV apparatus.I a.iii. .trenuous opposition to the author's proposal to base This paper deserves careful and reasonable analysis and shouldcoordination on a 90-percenit probability of flashover of the incite further exploration.switched jaw or hinged insulator assembly, producing less than I concur with Mr. Foti that agreement is desirable and neces-a 10-percent probability of switch open gap flashover. Because sary in the areas of:of the importance of and investments in 500-kV systems, thehighest degree of reliability should be designed into the network 1. Conditions for conducting EHV switch and bus supportand its parts. tests.

Experienice in our area has shown that the open gap of dis- 2. Method for determining switching surge withstand voltage.conniiectinig switches at the 69-kV and 161-kV levels are not im- I take this a step further to say that the statistical probabilitymune to flashover. Repeated flashover of open gaps has re- procedure should extend to impulse withstand and possibly toquired the installation of several lightning arresters to protect 60-cycle withstand, as well as to switching surge. The originalswitches normally in the open position. Most, if not all, of premise on which the probability approach was advocated wasthese flashovers have been caused by lightning impulses since based upon the wide dispersion of flashover values found inat this level the lightning impulse is considered a more severe positive switching surge testing. The proposed probabilityduty than the switching surge. At the 500-kV level, the switch- method does not correct this dispersion. It simply presents aing surge will impose a more serious stress on the open gap than mathematical means of accepting it.lightninig impulses, so the margin of safety should be increased 3. Definition of switching surge waveform to be used in labora-instead of decreased as proposed by MIr. Foti. tory testing.

In addition to these areas, it is also desirable to proceed asrapidly as possible to establish acceptable standards levels forR. F. Swoish and E. R. Perry (Allis-Chalmers Company, NMil- line-to-ground insulator behavior as a ratio of operating voltage.waukee, Wis.): The author is to be commended for his excellent Also desirable and economically mandatory is a standard systemwork in advance design and testing of disconnect switches for of establishing line-to-ground and switch-gap coordination.EHV application. He has presented some provoking ideas as To date, every prospective purchaser of apparatus in the 500-to the eventual need for standardization of testing and of test kV field has presented different specifications for establishingvalues at these extraordinarily high-voltage levels, and has suib- these two values.

stantiated his suggestions by test data. Referring to the second paragraph of Mr. Foti's conclusions,The author apparently is aware of the vertical-reach type of he states that the single-break switch has been selected for EHYTswitch which is utilized in Europe to some extent for high-voltage service, 500 kV and higher, because it offers better than averageand EHV lines. The high production cost of this type of switch switch open gap insulation coordination without requiring meanshas precluded its use in the United States. Recent development for controlling the insulator dielectric strength.work has improved the operating capability of this type of Extensive tests on double-gap switches of various gap dimen-switch and, at the same time, reduced the cost to where it is sions have demonstrated behavior as stable as that of single-gapbelow that of a single vertical-break disconnect switch with switches. Obviously, more total gap is needed for the doublethree insulators. This type of switch requires a special sub- break than is the case for the single break. However, randomstation design, different from that normally used by American dispersion of switching surge flashover behavior affects one typeutilities. However, the cost of this station for a vertical-reach of switch as much as the other.type of switch is substantially less than that of a conventional Referring to insulator withstand values shown in Table I,substation. Moreover, the substation space requirements may it is noted that these values for dry positive switching surgesbe reduced as much as 20 or 30 percent. vary over a range of 13 percent. Values for wet positive switch-Where it is necessary to utilize conventional substation de- ing surge vary over a range of 18 percent. In order to coordinatesigns, a center vertical-break switch, which gives the same a switch gap properly, the higher value must be taken, whichelectrical characteristics as the single vertical-break switch, means the switch gap, if coordinated with one of the insulatorsrequires less overhead room, and is less expensive; this switch having a lower value, might not meet the requirements if ahas been developed and completely tested. It will be operable different insulator were used. The presence of this problemon 500-kV systems within the next year. makes it desirable to provide means, such as configuring hard-The sinigle vertical-break switch has the capability of utilizing ware and applying insulator gaps, to reduce the range of flashoverresistors during the opening stroke which the center vertical- between different insulators. It appears obvious that this isbreak switch cannot do. However, there are many applications a desirable procedure for either type of switch. Since the usewhere resistors are not required, such as on circuit-breaker dis- of gaps limits the maximum flashover voltage of the insulator,connlect switches. Using a center vertical-break switch in these the switch designer can with more certainty coordinate theapplications reduces original costs and affords close tolerance switch gap with the line-to-ground insulation.over the switch blade during an operation, as mechanical linkages Figure 7 is a photograph of a phase-to-phase flashover of aare not required and the switch blade lengths have been reduced double-break switch. This is interesting in view of the fact.as much as 50 percent. that flashover tests performed phase-to-phase on two properlyThe two cited designs are presently on order for installations coordinated switch poles produced no such flashovers. Theseon two Eastern utility systems at EHY levels. It appears from tests were performed between two double-break switches, placedtests conducted on the center vertical-break switch that the data on 25-foot phase spacing with the blades open and the far-endpresented in this paper is applicable and would appear to be terminals of each switch pole grounded. A positive switching

878 IEEE. TRA\N,sACTIONs ON POWER APPARATUs AND SYSTEMS OCTOBER

surge at the withstand value of the insulator was imposed on one and the other outdoors. Different objects were tested; theirterminal, and a negative switching surge of the same value was 50-percent sparkover voltages were mostly between 1000 andimposed on the corresponding terminal of the other switch pole. 1600 kV. Values of 2000 kV were reached, nevertheless, onRepeated discharges with incremental raising of voltage resulted insulator strings for 735-kV lines. Objects supposed to havein flashovers, phase-to-phase. However, all such flashovers grounded elements surrounding them, such as tower windows,occurred between the terminal or switch contact hardware of were tested with dummy elements and some real ones whileone pole to a similar portion of the other pole. None of the other objects were tested with the maximum clearance kept,flashovers occurred through the blade gap system as shown in as seen in Figs. 13 through 16 of this discussion.Fig. 7.

It appears that the method of making the test illustrated inthe picture was to overinsulate two adjacent contact terminals,open the blades, ground one of the terminals, and impose voltageon the other one until a fla-shover occurred. Obviously, with suchan arrangement, the weakest gap seen is the first portion of thedouble-break switch gap. When this gap is broken down, theinext course of procedure is to successively flash from blade toblade and then to the grounded terminal of the other pole ofthe switch.

It would be interesting to determine:

1. If the gaps of the two poles shown were coordinated so thatin a normal test the minimum flashover of the switch gap wassufficienitly above the maximum flashover of the insulator fornormal application values for the insulator.2. What the metal-to-metal clearance was in each of the switchgaps and what the applied voltage was when the flashoveroccurred.3. Confirmation of the supposition that the total eniergy wasapplied on one pole with the other grounded and that the two.switch poles were not complete but consisted only of two bladesystems and two contact and terminal systems. Fig. 13. EHV switch, rated 525 kV, was tested outdoors in CESI

Again, I must congratulate Mr. Foti on an excellent presenta- laboratories, with switching surges dry and with positive po-tion, but must disagree upon the major premise for selecting a larity in the range of 1400 kVsingle-gap switch over a double-gap switch.

Gianguido Carrara (CESI, Milan, Italy): I wish to make someobservations on many points of the paper which deal with generaltesting methods because they are of great importance in evaluat-ing test resuilts.As far as standardizations, agreements, rules to be established,

etc., are concerned, I wish to draw attention to the work presentlybeing performed within Working Group 2 of the IEC TechnicalCommittee 42, "High-Voltage Measuring Techniques." ThisWorking Group is preparing a draft for the revision of IECPublication 60, "High-Voltage Measuring Techniques," whereswitching surges will be dealt with, condensing the experienceand thoughts of technicians of the whole world.A second major point refers to the energy output of the impulse

generator necessary to prevent the streamers, or partial dis-charges, from producirng voltage drops that are too extreme.Such voltage drops may prevent the streamers from being con-verted into disruptive discharges.When switching surges are involved, it is only the front

capacitance which prevents such voltage drops, supplying thecharges necessary to the streamer currents. Actually the dura-tion of the front of the wave to represent the switching surgesrequires such a high front resistance that the charges from thegenerator may not be supplied "in time."One can, however, see on the oscillograms if there are voltage

distortions severe enough to raise doubts about the validity ofthe results.At CESI, we have recently performed tests on rod-rod and rod-

plane gaps in the range of 1000- to 1200-kV positive polarity,135/3000-pus waves, with the minimum -1000-pF (picofarad)and maximum -6000-pF values of front capacitance available.The differences found were well within the 2- to 3-percenttolerance limits of such tests, even if some distortion appearedon the oscillograms. The matter is interesting also from thestandpoint of the discharge mechanism.Another important point is the comparison between indoor and

outdoor tests. It seems useless to enumerate all the advantagesof performing the tests indoors whenever possible. Only thelack of clearance should compel outdoor testing.No complete and organized research has been done at CESI

as yet. During the last year, however, records were kept of allinteresting results, making available at present about 20 pairsof results, each of which was obtained with identical conditions Fig. 14. The same switch as in Fig. 13 is shown but with testsprevailing, except that one test of the pair was performed indoors made in CESI indoor laboratories

1965 FOTI: EHV DISCONNECTING SWITCHES 879

,> ~ We conclude from these results that, as far as switching surgetests in our laboratory are concerned, we can expect no better

'~~~'~~ agreement between results than 4±5 percent. Since practicallythe samie agreement is obtained indoors as outdoors, indoor testsare as reliable as outdoor tests, the only precaution being torepresent as closely as possible the environmental conditions.If a check of indoors vs. outdoors is repeated a few times andshows an agreement as good as 45 percent, there is a good chancethat indoor tests will be as good as outdoor ones as far as clear-ances are concerned.

Arem Foti: As evidenced throughout my paper, the design andapplication of EHV disconnecting switches is fraught with manyperplexing problems and controversial issues. One of the primemotivating forces prompting me to prepare the paper was the

Fig. 15. A full-scale 735-kV tower for outdoor testing of insula- urge to get the issues out in the open, in the hope of stimulatingtor strings in CESI laboratories. Tests were conducted with a discussion of the problems and offering technical data aid inswitching surges, dry and wet, positive, wet negative, and resolving the controversies. I am delighted with the interest

polarity in the range of 2000 kV that has been shown, as evidenced by the number of writtendiscussions. I wish to thank the discussers for their efforts andfor their contributions.

Especially encouraging is the endorsement of the newly pro-posed philosophy for establishing switch open gap insulationcoordination, found in the discussions of Messrs. Williams,Killian, and DeSalvo.Almost invariably, the discussers agree with the expressed

need for standardization of tests and testing procedures forEHV disconnecting switches. Our agreement in this regardwas richly rewarded by the good news imparted by Mr. Uptonwhen, in his discussion, he informed us that the Air SwitchWorking Group of IEEE has made good progress toward thegoal of standardizing EHV tests and testing procedures.Mr. Williams raised a question concerning a difference in

wave shapes reported for Figs. 8 and 9. The front of the wave,or the time to reach crest voltage, appears to be more influentialin switching surge test values-the tail of the wave havingnegligible effect, especially if the time for the decay to half-crestvoltage is in the order of magnitude reported. Mr. Williams'comments concerning RIV are well taken. The data given inFig. 11 of my paper tend to show that all insulator types testedyielded good results.

Several discussers: namely, Messrs. Williams, Miltoii, Swoishand Perry, have chastised me slightly for passing over ratherlightly the vertical-reach or pantograph switch. I am sorry tohave created this impression, but the company I represent madean earnest effort several years ago to market a vertical-reachswitch for EHV applications; there were no "takers"; conse-quently, the project was abandoned. Changing times and shift-ing demands could conceivably cause reconsideration as to usinigthe vertical-reach switch. At the moment, however, the com-ments expressed in the paper constitute an accurate appraisal ofthe situation to date.From the standpoint of insulation coordination of the switch

open gap of a vertical-reach or pantograph switch, I would cau-tion those who oppose the proposed 90-percent to 10-percent newphilosophy of coordination and advocate the pantograph switchthat said switch, in the open position, may not behave well underswitching surge test conditions. If the jaw conductor of suchan open switch were energized and the blade-end grounded,the relatively large space occupied by the collapsed pantograph

Fig. 16. The window of the tower of Fig. 15 reproduced indoors would make it behave like a ground plane. Tests have provedwith dummy elements for the same test as made previously that, for a given spacing, rod-to-ground planes yield lower

withstand and flashover values than the same spacing, rod-to-rod.In discussions by Messrs. Williams, Upton, DeSaivo, Swoish,

and Perry, comments were offered concerning the vertical-opening center-break switch generally, and some mention was

The generallpicture is that outdoor tests may give either lower made of its hydraulic operation. My views tend to be in agree-or higher results than corresponding indoor tests. On the ment with Mr. Upton's, in that I consider this type of switch to beaverage, outdoor tests resulted in slightly higher results (less more appropriate for voltage levels higher than 500 kV.than 2 percent); the absolute values of the differences were all Hydraulic operation of outdoor switches may have the ad-less than 9 percent (44.5 percent with respect to the average). vantage of operating more than one switch as mentioned by Mr.To allow a comparison, during the same period, 40 more pairs Swoish and Mr. Perry. However, I find it difficult to conceive

of tests were recorded, but with all tests performed indoors; that any hydraulic operator would be more economical than thethat is, in identical conditions in every respect. In this case, normal mechanical operating mechanism, especially if an EHVthe absolute values of the differences were less than 10 percent disconnecting switch user is perfectly satisfied with a manual-(45 percent with respect to the average). operating mechanism, arid is perfectly contented to have indi-

880 IEEE TRANSACTIONS ON POWER APPARATUS AND SYSTEMS OCTOBER

vidual single-phase operationi. It should also be mentioned is a possible "yes" but, in some cases and for specific types ofthat the hydraulic system is vulnerable to damage, especially voltage stress, protective gaps may be required to control theif a line-to-ground fault were to occur in the immediate vicinity line-to-ground dielectric strength in order not to increase theof the hydraulic cylinder or the insulator column supporting it. switch open gap distance.A flashover of the "oil port" insulator column and the follow I would like to discourage Mr. DeSalvo in applying confidencepower current would need only to damage the insulator to the limits to the switch open gap coordination on the basis that itsmall extent of developing a crack into the porthole and the is too precise and not realistic for the relatively inexact specimenhydraulic fluid would be lost. Or, a similar-type flashover that is under consideration. The confidence limit study iscould conceivably burn a hole in the hydraulic cylinder itself, fine from a theoretical standpoint. It is known that confidencebecause it is located at the energized part of the switch. This, limits are large when the number of trials is small. This can betoo, would result in loss of hydraulic fluid and, consequently, offset in a test program by increasing the number of trials at arender the switch inoperable. given voltage level from 10, for example, to 20. My point is

Answering other questions as raised by the individual discus- that the confidence should be ground into the testing proceduresers, let me first cover those of Mr. Burdeshaw, relative to and should not be projected into the specification for switchFig. 7 of the paper. An honest effort was made to have the gaps open gap insulation coordination.coordinated as stated by Mr. Burdeshaw. The metal-to-metal I wish to thank Dr. Carrara for giving the excellent comparativeclearance of the switch gaps was slightly more than 100 inches, test data between indoor and outdoor EHV tests. His labora-with 87 inches metal-to-metal clearance between the tips of the tory is no doubt one of the few places in the world where suchadjacent phase switch blades. The applied voltage at which a direct comparison could be drawn. It is interesting to noteflashover occurred was in the order of magnitude of 1750-kV the tolerance of 41//2 percent and 45 percent for outdoor andpositive-polarity switching surge under dry conditions. Mr. indoor tests, respectively, covering the agreement that could beBurdeshaw's other assumptions concerning application of voltage expected in test results, obtained for identical tests at differentand configuration of tested blades are correct. times. Perhaps this should be a lesson for tolerance in EHVThe implied question by Mr. Swoish and Mr. Perry regardiing equipment specifications, and in the ability to precisely meet

appropriateiiess of using the single-break switch data of the the requirements during tests, or to comply precisely with cal-paper for their proposed center-break vertical opening switch culated theoretical probabilities. Needless to say, the materialis answered in the affirmative. presented by Dr. Carrara adds confidence to the validity of testMr. Milton, when speaking of the open gap of a 500-kV switch values published in my paper, since all the reported tests were

designed in accordance with the philosophies discussed in the conducted indoors at the CESI laboratory where he has demon-paper, asked if the tests upon which the conclusions were reached strated good agreement between indoor and outdoor tests.were made on extrapolated models. The answer is, "No." I conclude my remarks by expressing a debt of gratitude toThe tests were conducted on actual-size switches or switch parts. those who participated in discussions. A certain sense of satis-

Mr. DeSalvo called attention to the fact that Tables I and II faction is felt because the paper engendered these discussionsused the 90-percent flashover voltage value for pin-and-cap and because we can expose and amicably discuss controversialinsulators. He asked if a larger gap would be required had the EHV issues, leading to a likelihood that our problems and issuescomparison been made with post-type insulators. The answer will be resolved in the not-too-distant future.

500-kV Switch Design and TestsPayton C. Mayo, Senior Member IEEE

Abstract: A discussion of the development and test of 500-kV resters. In order to fulfill this demiand, manufacturersair break switches is presented. An explanation of the per- are busy designing, testing, and manufacturing EHV appara-formance requirements of extra-high-voltage switches and thereasons they differ from lower voltage switches is included. Pre- tus. This paper recites the exp)eriences and knowledgeliminary research tests and switch configurations are described, gained by one manufacturer in designing and testing 500-kVpresenting the problems of design and explaining solutions. air break switches.Tests of the switch installed on cap and pin and post insulators An air break switch provides an insulating air gap in a cir-are included for 60-cycle, impulse, and switching surge. The cuit so that lines or other apparatus may be isolated or con-gap tests described allow switch gap-insulator comparisons. Ct soperationalor requi Imsy be depen-Switching surge phase-to-phase test results are included, in addi- nected as operational needs require. It must 1) be depend-tion to the results of switching surge tests made of the switch ably operable to open or close under sl)ecified conditions, 2)when used to connnect two utility systems together. provide sufficient air gal) to assure that flashover strength

is greater than flashover strength to ground, 3) withstandAn unprecedented expansion of facilities for transmitting rated continuous and short circuit currents without damage,large blocks of power over great distances is prevailing. To and 4) perform at rated voltage without emitting excessivetransmit this power economically, electrical utilities are find- corona or radio influence voltage.ing that they must turn to extra-high-vroltag;e (EHV) trans- Air break switches of lower voltag;e ratings have been de-mission. Therefore, the demand is increasing for EHY air signed to standards as established by the National Electricalbreak switches, breakiers, transformers, and lightning ar- Manufacturers Association (NEMA) and the American________________________________________ Standards Association (ASA). These standards do notPaper 31 TP 65-84, recommended and approved by the IEEE necessarily provide assurance that an open switch will alwaysSwitchgear Committee of the IEEE Power Group for presentation flash over the insulators to ground instead of across the openat the IEEE Winter Power Meeting, New York, N. Y., ,January 31-February 5, 1965. Manuscript submitted November 4, 1964; gap. It is desirable to establish coordination to assure flash-made available for printing December 5, 1964. over of the insulator before flashover of the gap. The problemP. C. MAYO iS with Southern States, Inc., Hampton, Ga. in doing, so is economic, and operating exl)erience does not

IEEE TRANSACTlONS ON POWER APPARATUS AND SYSTEMS VOL. PAS-84, NO. 10 OCTOBER 19}65