reliability of lightning resistant

8/13/2019 Reliability of Lightning Resistant

1/6

Reliability of Lightning ResistantOverhead Distribution LinesLeon M. Tolbert, Member IEEE Jeff T. Cleveland, Member IEEE and Lynn J. Degenhardt, Member IEEE

RFtrocr n assessment of the 32 year historical reliabilityof the 13.8 kV electrical distribution system at the Oak RidgeNational Laboratory ORNL) in Tennessee has yielded severalconclusions useful in the planning of industrial power systems.The system configuration at ORNL has essentially remainedunchanged in the last 32 years which allows a meaningfulcomparison of reliability trends for the plant s eight overheaddistribution lines, two of which were built in the 1960%withlightning resistant construction techniques. Meticulous record sindicating the cause, duration, and location of 135 electricoutages in the plant s distribution system have allowed areliability assessment to be performed. The assessment clearlyshows Bow differences in voltage construction class, length, age,and maximum elevation above a reference elevation inffuencethe reliability of overhead power distribution lines.Comparisons are also made between the ORNL historical dataand predicted failure rates from ANSI and IEEE industrysurveys.

I. INTRODUCTIONIGHTNING continues to be the major c use of outagesL n overhead power distribution lines. Through

laboratory testing and field observations and measurements,the properties of a lightning stroke and its effects on electricaldistribution system components are well-understoodphenomena [l] Thispaper presents a compilation of 32 yearsof historical records for outage causes, duration, and locationsfor eight distribution feeders at the Oak Ridge NationalLaboratory ORNL). Due to the limited growth of ORNL, thenumber, ength and location of the 13.8 kV overhead lineshave remained the sam e between 1960 and 1992. Except fornoted differences (voltage construction class, length, age, andmaximum elevation above a reference level), other factorswhich couId influence the reliabifity of an overhead line haveretnained nearly the same. This allowed a meaningfulreliabiiity study to be performed on the entire ORNT.,electrical distribution system 121.This paper discusses the main findings of the reliabilityassessment as it relates to lightning resistant overhead lineconstmction techniques and offers a simple and cost effectivemethod to reduce lighting caused outages. In addition,comparisons are made between the failure rates and causesexperienced at ORNL and those in industry surveys. Wherelarge discrepancies exist between survey data and experiencesat ORNL, evidence is presented to explain the differences

between O N S distribution system and those typical oindustry.n.DESCRIPTION F ELECTRICAL DISTRIBUTIONYSTEM

Electrical power at the Oak Ridge National Laboratory( O W ) n Tennessee is supplied by a single 161kV to 13.8kV primary substation. The average electrical load is 19MW. Eight radial 13.8 kV overhead lines originate from thesubstation. The ORNL 13.8 kV distribution system iscomprised of six facility 13.8 kV switchgear equipment, 37padmounted unit substations and transformers, 46 polemounted transformer stations, 1 mile of underground cable,and approximately 20 miles of open overhead lines. The 13.8kV distribution system covers 10 squaremiles of service area.The lines rangein age from 23 to 42 years.The ORNL. 13.8 kV system is a wye system which isgrounded through a 4 ohm resistor at the primary substationto limit ground fault currents. Except for limited shortlengths, all three phase conductors are installed on each ofthe overhead lines. Autom atic reclosers at the primarysubstation are utilized on some of the overhead lines o limitdowntime. No in-line reclosers or fuses are installed in theoverhead lines, except for feeder 3 which has a single in-linerecloser. For purposes of the assessment, a succesfid recloseoperation is st i l l treated as an outage because facilityoperations and processes are shutdown, though onlymomentarily.The eight overhead lines have various lengths and followdifferent terrains in serving their electrical loads. Except forlimited short lengths, all of the overhead lines use separateright-of-ways. Some overhead lines in the system areinstalled on high ridges which are predominant in the region,exposing and making them vulnerable to Iightning . Due tothe ruralEast Tennessee location, insulators are not exposedto industrial pollution or salt water. The overhead lines havebeen moderately maintained, equipment has been replaced asneeded, and nominal 50 ft right-of-ways have been keptmoderately clear.The overhead lines use 50 to 75 ft tall woad poleconstruction and utilize overhead ground wires to interceptdirect lightning strokes. Typically at eveiy fourth pole, theoverhead ground wire is bonded to a 4 AWG copper wirewhich extends down the pole to a pole butt-wrapped ground

~ ~wire; no buried coun poise ground wires are used.


2/6

DISCLAIMERPortions of this document may be illegiblein electronic imag e products. Ima ges areproduced from th e b est available originaldocument .


3/6

Distribution type lightning arresters are placed at dead-endand angle structures, at pole mounted W o r m e r locations,and at high points on the overhead line. Station classlightning arresters are used to protect underground cableruns pad mounted switchgear, and unit substationtransformers.Resistance to earth of each pole ground is typically 15 ohmsor less. At higher elevations in the system, resistance to earthis substantially greater than 15 ohms, especially during thedry sum mer months. At these high points, ground rods weredriven and bonded to the pole grounding systems in the1960's in an attempt to decrease lightning outages. The seattempts were only partially successful n lowering the outagerate.From a surge protection standpoint, the variety of polestructures used (in-line, corner, angle, deadend, etc.) and thevariety of insulators and hardware used does not allow each13.8 kV overhead line to be categorized with a uniformimpulse flashover rating (170kV etc.) or a numerical BILvoltage class (95 kV BIL; etc.). For simplicity purposes inthe analysis, each overhead line was categorized with anominal voltage construction class (15 kV, 4 kV, r 69 kv .Six of the eight overhead lines (feeders 1 through 6) werebuilt with typical REA Standard horizontal wood crossarmconstruction utilizing single ANSI Class 55-5 porcelain pininsulators (nominal 15 kV nsulation). The shield angle ofthe overhead ground wire to the phase conductors is typically45 degrees.One overhead line (feeder 7)was built with transmissiontype wood pole construction because the line extended to aresearch facility which was to have generated e lectrical powerto feed back into the grid. Pole str uc tu m of this line are ofdurable wood cross nn construction which utilize doubleANSI 52-3 porcelain suspension insulators to support the

conductors (nominal 34 kV nsulation). The shield ang le othe overhead ground wire to the phase conductors for feederis typically 30 degrees.In 1969, an overhead line (feeder 8) was intentionally buiwith ligh tning resistant constxuction in an attempt to reduclightning caused outages. Pole structures of the line havphase over phase 24-inch long fiberglass suspension bracketwith double ANSI 52-3 porcelain suspension insulators tsupport the conductors (nominal 69 kV insulation). Thshield angle of the overhead ground wir to the phasconductors for feeder 8 is typically 30 degrees.

III. FAILUREANALYSISA. System OperatingHistory

For the purposes of t h i s paper, an electrical outagedefined as the unplanned loss of power to a load. Thcondition is also commonly referred to as a forced outage oa failure of the power system component under study, (ithis case he overhead distribution lines) [3]. Instanas otransient power d resulting in a momentardegradation of power quality (e.g., voltage sags and surgeswere not recorded and, s a result,were not considered in thassessment.A review of the historical data showed each outage co dd bcategorized as being due to one of five distinct failureinitiating events: weather-related phenomena (ligh tningwind, etc.), animal contact, equipment failure (insulatiobreakdowq loose connections, etc.), human error (usuallyswitching errors), or unknown factors. The failure data wacompiled for each of the eight 13.8 kV feeders and ipresented in Table I, along with pertinent informatioregarding feeder construction, elevation, length, and age.

TABLE IFEEDER CHARACTERISTICS ND OUTAGE CAUSES

Total

7 56 )

28 (21 )18 (13%)7 (5 )

2


4/6

A key finding of the failure analysis is that w eather-relatedevents account for over half (56 ) of the feeder outagesrecorded. Fifty-seven of the 76 weather-related outages wereattributed to lightning. Insulation breakdown damage due tolightning is also suspected in at least a dozen of theequipment failures observed. The data indicates overheadlineswhich pass over high terrain are less reliable because ofthe greater exposure to lightning. For example, feeder 3 hadthe most recorded outages (48), of which two-thirds were dueto weather-related events; this feeder is also the highest lineon the plant site, rising to an elevation of 450 ft above thereferencevalley elevation. Overhead lines that are longer andto which more substations and equipment are attached werealso observed to be less reliable (more exposure to lightningand more equipment to fail . The age of the line does notappear to significantly lessen its reliabiity as long asadequate maintenance is performed; none of the lines havehad a notable increase in the frequency of outages as the lineshave aged.Aswould be expected, the empirical data presented in Ta bleI confirms the two overhead lines which have been insulatedto a higher level (34 or 69 kv) have significantly betterreliability records th n those utilizing 15 kV classconstruction. Feeder 7 (insulated to 34 kv) and feeder 8(insulated to 69 kV) have bad only 3 outages each over their32 and 23 year life spans, respectively. These lines followsimilar terrain and are comparable in length and ag e to the 15kV class lines,yet they have a combined failure rate of 0.22failures per year versus 4.32 failures per year for therem ainin g feeders.

B. Comparison to Survey DataAlthough the extensive historical data points to lightninand lightning-induced equipment failures as majocontributors to the ORNL feeder outages, it is also useful testablish whether the frequency and nature of these outageare typical of industry. To accomplish this the failure cause

of all ORNL feeders were re-tabulated to match the categorieof failure-initiating causes for open wire installations apublished in ANSI/IEEE Standard 493-1990 p] . Thstandard presents the results of an I sponsored reliabilitsurvey of industrial plants completed in 1972. Table Icompares the data obtained from this survejr to that compilefrom theORNL reconls.The comparison between he ORNL istorical data and thIEEE survey data indicates that, as a percentage of totaoutages, transient overvoltage disturbances at ORNL artwice the industrial average. Based upon the empirical datpresented, t h is discrepancy is due to a heightenevulnerability of the plant's overhead distribution system tolightning. This heory is also supported by the fact that thregion's relatively high isokerauaic level subjects thsurrounding area to approximately 60 hunderstorm days peyear. At an isokeraunic level of 60 hunderstorm days peyear, one can estimate two lightning strokes per m ile of linper year 141. Insutation breakdown is also double thindustrial average. Again, lightning damage to conductoinsulators and lightning arresters is suspected for thianomaly.Explanations for the differences in the other categories ar

TABLE IICOMPARISONOF FAILURE CAUSES blDUSnUAL AVERAGEVS.o m CTUAL

IEEE Survey 131 All FeedersFailure-Initiating Cause Industrial Average at ORNL

1 Transien t overvoltage disturbances (lightning, 26% 47%

2 Overheating 21% 0%

4 Mechanical breaking, cracking, loosening, abrading, 7 7

switching surges, arcing ground fault in ungrounded system)

3 Other insulation breakdown 8 16%

or deforming ofstatic or structural parts5 Mechanically caused damage from foreign 10% 6

source (digging, ve h id a r accident, trees, etc.)6 Shorting by tools or metal objects 14% 77 Shortingby birds, snakes, rodents, etc. 3yo 4%8 Other (human error,unknown ause, etc.) 11% 13%

3


5/6

also postulated. Overheating Eailures are less than theindustrial average because ORNL does not overloadtransformers or conductors to the extent that general industrydoes. Failures resulting from vehicula r accident, vandalism,digging, and shorting by objects are less likely because of therestricted access to the ORNL ite. Other failure causes (e.g.,mechanical part failure, animal contact, and human errorwere ound to be sim ilar to the industrial averages.Standard 493-1990 also presents failure rates for twosubcl sses of open wire installations (0-15 kV and above 15kv) 3]. As c n be seen from Fig. 1, failure rates for the 15kV cIass ORNL feeders are over four times the industryaverage7 which is indicative of the substantially higherpercentage of lightning-induced failures noted in Table II.However, the feeders employing lightning-resistantconstruction exhibit outage rates reasonably close to theindustry average for the above 15 kV subclass. Thiscomparison provides additional evidence supporting theeffectiveness of these techniques in areas susceptible tofrequent lightning.

T 0.411g 0.4>a 0.3

0.2I90 0.1

015 kV andBelow Above 15kV

Voltage Construction Class

Fig. 1. Outagesper Mileyear Based on Insulation Level

N IGHTNINGESISTANT CONSTRUCITONThe term lightning resistant has been arbitrarily defined

as the use of high nsulation (34 kV or 69 kV) hardware whenapplied on a lower voltage (13.8 kV, 4.16 kV7 etc.) overheadline as a means of appreciably decreasing the lightningoutage and recloser action rate. The increased insulation isapplied in conjunction With overhead ground wires lightningarres ters and pole structure grounding. In effect, theconstruction method brings overhead distribution lines up tosubtransmission lightning protection standards Throughjudicious application, the construction can be applied tosections of distribution lines which have greater lightningexposure in order to reduceweather rela ted outages.

On typical 15 kV insulated line construction lightninflashovers often cause 60 cycle power follow and feeder tripWith the higher insulation construction outage rates arreduced by limiting the number of flashovers and thresultant power follow which causes an overcurrent device totrip. This allows lightning arresters to perform their duty odissipating lightning energy to earth. The number of recloseactions and their resultant momentary outages are alsreduced. This s beneficial for critical facilities and processewhich c nnot tolerate even momentary outages. nadditional benefit is that outages due to animal contact aralso reduced because of the greater distance from phasconductor to ground on pole structu res.Distribution line equipment to increase line insulatiovalues are off he shelf items an d proven technology. Newlightning resistant construction typicaUy utilizes horizontaline posts, fiberglass standoff brackets or any other methodwhich wo dd increase the insulation value. The eplacemenof standard pin insulators with line post insulators of greateflashover vd ue is an effective means to retrofit existing woodcrossarm construction. The doubling and tripling of deadendand suspension insulators is also a means of increasinflashover values on existing angle and dead-end structuresCurrent fiberglass, polymer, and epoxy technologies providan affordable means to increase line insulation.

V. CONCLUSIONWhile the use of increased insulation levels to redulightning flashovers and the resultant outages on overheaddistribution lines has been thoroughly tested anddemonstrated in laboratory and experimental tests [5], longterm historid field data has positively demonstrated that the

use of lightning resistant construc tion c n greatly reduceoutages. Field use at ORNL has shown that in areas whichare vulnerable to lightning, the use of increased insulationand a smaller shielding angle is an impressive and coseffective means to appreciably increase the reliability ofoverhead distribution lines.This reliability study clearly illustrates that the insulationrequirements for high-reliability distribution feeders shouldbe determined not by the 60 Hz operating voltage but ratheby withstand requirements for the lightning transients oother high voltage transients that are mpressed upon the lineElectrical equipment (switchgear, insulators, transformers

cables, etc.) have a reserve BE evel or flashover value) tohandle momentary overvoltages, and by increasing thareserve, the service reliability is appreciably increased.As the electrical industry gradually moves away fromstandard wood crossarm construction and moves toward morefiberglass, polymer and epoxy construction, increasedinsulation methods c nbe applied as part of new constructionor aspartof an upgrade or replacement effort. In consideringnew or upgraded overhead line construction, the incrementaincreased cost of the higher insulation equipment is nproportion to the total costs of construction (labor, capita4


6/6

equipment, cables, electric poles, right-of-way acquisition),Its cost effectiveness varies with the application and theconditions to which it is be applied. Economic benefitsinclude increased electrical service reliability and its inherentability to keep manufacturing processes and critical loads inservice. Other more direct benefits include less repair ofoverhead distribution lines, which c n have a significantreduction in maintenance cost due to less replacementmaterials and a large reduction in overtime hours formaintenance crews.

ACKNOWLEDGMENTThe authors wish to thank Gerald B. Young of theElectrical Maintenance Department at the Oak Ridge

National Laboratory. His meticulous record keeping forover 30years allowed this analysis o be made.REFERENCES

(11 C. F. Wagner, G. D. McCann, and J. M. Clayton, LightningPhenomena , Surge Protection of Power Systems WestinghouseElectric Corporation,September, 971, pages 1 19.[2] L. J. Degenhardt A Study of the ORNLElectrical Distribution[3] IEEE Standard 493-1990, I Recommended Practice or theDesign of Reliable Industrial nd Commercial Power Systems[4] A. C. Monteith, E. L. Harder, and J. M. layton, Line DesignBased Upon Direct Strokes , Electrical Tmnsmission andDistribution Refirence Book. Westinghouse ElectricCorporation, 1964, page 585.[SI A. J. Eriksson, M. F. Strinsfellow, and D. V. Meal, Lightning-Induced Overvoltages on Overhead Distribution Lines , ZEEE

Transactions o Power Appamt~~nd Systems Vol. PAS-101,No. 4, April 1982.

System , ORNuCF-9u268, November, 1992.1990,pages 54,75,204.

Leon M. Tolbert (S SS-MYl) received the B.E.E. degree in 1989and theM.S. egree in 1991, both in electrical engineering fiom theGeorgia Institute of Technology, Atlanta.He worked as a co-op in 1989 for International Business Machinesin Lexington, Kentucky, and in 1990 for bascoServicesinAtlanta.He is currently employed as a design engineer by h4artin MariettaEnergy Systems at the Oak Ridge National Laboratory in TennesseeHe is pursuing a Ph.D. degree in electrical engineering, and hisfields of interest include power quality and control of electricmachines.Mr. Tolbert was the recipient of a Prize Paper Award by theIndustrial Drives Committee at the I Industrial ApplicationsSociety 1991 AnnualMeeting.Jeffrey T. Cleveland (M 83) received the B.S.E.E. and M.S.LE.degrees from the University of Tennessee, Knoxville, in 1983 and1990, respectively.He has been employed as an electrical engineer at Oak RidgeNational Laboratory since 1983,where he has been responsible forthe design of electrical power distribution and telecommunicationsystems for several plant facilities. He is currently a departmenhead in the Instrument and ElectricalEngineeringgroup at ORNL.Mr. leveland is a Registered Professional Engineer in the state ofTennessee.Lynn J Degenhardt (hf94) eceived the B.S.E.E. degree in 1969from he University of Missouri-RollaHe is an engineering design specialist working fw MartinMarietta Energy Systems at the Oak Ridge National Laboratory inTennessee. His experience includes high voltage substation andelectrical distribution system design and operation.He joined theOak Ridge National Laboratory in 1974. Prior to 1974, he wasemployed by McDonnell-Douglas, where he worked on the Skylabspaceproject.Mr.Degenhardt is a Registered Professional Engineer in the stateof Tennessee.

DISCLAIMERThis report was prepared as an a m u n t of work sponsored by an agency of the United State sGovernment. Neither the United S tates Government nor any agency thereof, nor any of theiremployees, makes any warranty, express or implied, or assumes any legal liability or responsi-bility for the accuracy, completeness, or usefulness of any information, apparatus, product, orprocess disclosed, or represents that its use would not infringe privately owned rights. Refer-ence herein to any specific commercial product, process, or service by trade name, trademark,manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recom-mendation, or favoring by the United States Government or any agency thereof. The viewsand opinions of authors expressed herein do not necessarily state or reflect those of theUnited State s Government or any agency thereof.

5

reliability of lightning resistant

Documents