ERA REPORT 'J7 - () "33 Earthing Sn ltlli ons , Sta ncla r rls . Saf ct y X r;o() {1 Prnct ice PP 1 . 2 . l - 3 . 2. 15

EXPERIMENTAL METHODS FOR THE EXAMINATlON OF SUBSTATlON EARTHlNG SYSTEMS

Th. Aschwandcn, R. Briiunlich

FKH, Fachkommission für Hochspannungsfragcn, Switzcrland

AUTHOR BIOGRAPHICAL NOTES

Thomas H. Asehwandcn reeeived his Dipl..lng. and Ph.D. degrees in eleelrieal engincering from Ihe Swiss Federal Inslilule of Teehnology (ETH) Zürieh in 1978 and 1985, respeelively. He joined Ihe Weber Researeh Inslilule, Polyteehnie Universily of New York, in 1985 and worked lhere as a researeh assoeiale in Ihe field of eleelrophysies and pulsed power leehnology. In 1987 he beeamc senior assislanllo Ihe high vollage engineering group of ETH. Sinee 1990 he has been Ihe head of Ihe high voltage researeh eommission (FKHl of Ihe Swiss power ulililies and induslry. The researeh aetivities of Dr. Asehwanden include eleetrophysies, high voltage teehnology, eleetromagnetie eom· patibility, pulsed power, dieleetrie materials and insulation diagnosis, lightning proleetion , on-site tesling of high vollage apparatus and systems. He is a member of IEEE, CIGRE and Ihe Swiss Eleetroteehnical Assoeiation (SEVl.

Rcinhold H, Braun\ieh reeeived his Dipl.-Ing. degree in eleetrieal engineering in 1982 and his Ph .D. in 1993 from the Swiss Federal Institute of Teehnology (ETHl, Zürieh. After a short period as a development engineer with BBe Brown Boveri & Cie., he joined the high voltage engineering group ofETH in 1983. Sinee 1991 he has been with the high voltage researeh eommission (FKHl in Zürieh, where he works in the field of on-site high voltage testing. He is a member of IEEE and the Swiss Eleetroteehnieal Assoeiation (SEVl and therein a member ofthe Power Soeiety (ETGl.

ABSTRACT

Experimenlal methods for the judgemeJ1l of large earthing systems of high voltage subslalions including AC faull eurrent injeetion are deseribed. For the verifieation of the requiremenls for personal safely il is usually suffieienl lo assess the earthing voltage of a substation as well as adequately seleeled samples oftoueh vollages. An important requirement for reliable measuremenls of loueh voltages in operaling subslations is Ihe diserimination against disturbing signals originaling from the line eurrenls. It is shown, lhat interfering signals may easily be suppressed by using as}'nehronous fault current injection with a frequenc}' different from line frequenc)' (50 Hz) . Evaluation of Ihe current dislribution in the event of an earth fault provides further insighl into the condition and state of large and complcx earthing systems.

Introduction

Experimental examinalions of earthing systems serve first of all as a means to test personal and teeh­nieal safety in ease of an earth fault. The goal of so ealled "earthing measurements" is Ihe judgcment of the efficieney of an earthing system . Within the area of inOuenee of eleetrie installations, no immoderate values of earth potential differenees or earth eurrents, whieh might endanger living beings or teehnieal installations, must oeeur. For the ease of an earth fault, the admissible toueh voltages on aeeessible melal objeels are limited . In Europe, these limits are speeified in national standards usually laid down as a voltage-time-eurve (see Figure I l. These definitions are based on the sensitivity of the human body to AC eurrents; the limiting values differ somewhat from eountry to eountry [8 ... 12] .

At the moment, efforts are being made to harmonise the rules for earthing installations as part of Ihe CENELEC slandard "Power illsta/lalions exceedillg I kV AC". A preliminary version is already available [I] .

In addition to safety malters, the judgement of an earthing situation should also inelude operational viewpoints. Here, examinations on mutual disturbanees between different adjaeent installalions and

3.2. 1

ovcrload oI" cunductors 111 Sf1t:'clcll Operil[ IOllal <.;lIucltinns arc in Ihe cCl1lrt:' li ' Intere st. Anothcr 1I11portant

qucstion 15 Ihe minimisation of encrgy loss during opcration duc lo cJo5ed (carthingJ currcnl loops.

hlr a (omf1lctc judgcmcllI o r iI "i uhSlalion carthing. not only the carlhing "iys tcm bu! .11 50 Ihe silualiOIl

of Ihe grid , Ihe Iypes of po,,"er Ilnes and eanh eonneelions la olher leehni ca l ,nslallalians in Ihe vieinily of Ihe subslalion have la be eonsidered. 11 has lO be kepl in m,nd Iha I usual !Oueh vol!age examinalions al eleelrie power inslallalions ean only be earried oul by SpOI eheeking. The seleelion of Ihe lesl poinls presumes an exael knowledge of Ihe whole silUalion of Ihe examined subslalion .

The measuring melhods presented here are reslrieled 10 eanhing eondilions al power frequeney (50 Hz, 60 Hz ar Iraelion syslems al 16.7 Hz). More preeisely, Ihe eanhing inslallalions are examined wilh regard 10 eanh faul! silualions Ihal are simulaled by an artifieial, large earthing eurrenl loop (see Figure 2) . This experimenlal viewpoinl is also Ihe basis for Ihe design of an eanhing syslem, so Iha I dimensioning ean be verified direelly. In ease of an earth faull, lemporarily, a high eurrenl Dows Ihrough Ihe earthing syslem inlO Ihe soi!. The dislribulion of Ihe faull eurrenls eauses a (Iransilory) rise in pOlenlial of Ihe earthing syslem, whieh is responsible for Ihe slep and loueh vollages lo be examined and judged.

An importanl part of an earthing syslem analysis is Ihe evaluation and inlerpretation of Ihe measured values. Unexpeeledly high ar ineonsistent loueh vollage values ean often noI readily be explained. However, knowing further dala and inforrnalion aboul Ihe examined syslem (i.e . geometrieal dispo­sition of the earthing grid, potenlial and eurrent distribution) helps 10 eome to a eonelusion and lo suggest remedial measures.

Finally, Ihe experimental examination of an earthing system should generally provide inforrnalion an weak points and data for a further oplimisation ofthe earthing arrangement.

2 Injection of earth fault currcnt - formation of earth fault currcnt loop

The experimenlal examinalion of Ihe pOlelllial and eurrent distribulion of an earthing syslem is essen­lially based an Ihe eurrent-vollage-method [li, [2J . For this purpose, a defined earth faul! eurrent lE is injeeted into a defined eanh faull eurrenl loop, which eomprises Ihe earthing system 10 be examined . The experimental earth faull eurrenl is generaled by a single-phase AC eurrenl souree ar by an impulse eurrent souree. Usually Ihe eurrenl amplitudes used for fault eurrenl simulalion are limited by eireuil impedanee and souree: Iyp ieal values in Ihe 220kV- ar 400kV-grid are 100 - 200 A (AC).

for large earthing syslems Ih e eanh faull eurrenl loop is usually fonned by an o\'erhead line or a eable whieh has lo be laken oul or opera lion and earthed in a neighbouring subslalion (auxiliary or referenee earth, see figure J). Generally, Ihe examined substations remain in operalion. and Ihererore, inlerferenee from serv iee eurrenls (i.e. 50 Hz) and lheir harmonies have lo be expeeled in all earthing system measurements. The distanee bel\Veen th e l\Vo substations used lO rorm Ih e earth eurrenl loop has lo be suffieiently large (more Ihan I km), so as to exclude any mUlual inOuenee of Ihe earthing syslems. The phase eonduelors ar Ihe power line used to feed the experimenlal earth fault eurrent are often eonneeled in parallel to reduee Ihe impedanee ofthe loop (zero sequenee impedanee) .

The potential of the eanhing syslem under lesl ri ses to the earthing voltage Ul: by Ihe aelion of injeeting the earth fault eurrent (see figure 2). The earth potential deereases lo 10\Ver values wilh inereasing distanee from the boundary of Ihe earthing system. Any objeels \\'ithin the 'hot zone' around the substation area eneounler a pOlential gradient, and therefore, sho\\' a voltage differenee wilh respeet lo the surroundings. With appropriale voltmeters, these voltage differences ean be measured as loueh or step voltages'.

The experimental earth eurrents eireulaling from one substation to Ihe olher are nat only Oowing Ihrough the soil bul also Ihrough eable shea lhs and ground wires of overhead lines, as already men-

A 'hal zone' is also eSlablished around Ihe subslalion used for auxiliary (referencc) eanhing. Therefore, earthing mcasurements can be pcrformcd on both substations fonning the expcrimcntal carth fault currcnl loop (Figure J).

3.2.2

IHJlH.:d 111 lbe IlllroduCIIOn . Othcr extcndcu COI1UUC tUlg structllrcs \uc h a ~ tuhc duct s, ra ilway line ... ur rclccam lines conducl a ran a r (he rault current as wel! . The assessmcll t ar Ihcse conduclor-bound

cu rren( s is a l50 an integral part or an earthing syslcm examina(ion

Mcasured results (!Oueh and step vo ltages, eanh eonduetor eurrents) are subsequently linearly extrapo· lated to the maximum eanh fault eurrent o f the examined installation . This rna xi mum eanh fault eurrent o f a high voltage substation with so lidl y eanhed neutra ls is usually pradueed an the oeeasion of a line·to-eanh fault an a bus bar and is dependent an the state of the grid .

For the inJeetion of the eanh fault eUITent lE a eurrent souree wit h an ampaeity of a fe w tens to a few hundreds of amperes is required for large eanhing systems of high voltage substation s ar power plants. The measurements deseribed in this paper were perforrned e ither with a 100 k V A diese l generator ar with an eleetranie frequeney eonvener with a power of 50 kVA . For proper matehing with the im­pedanee o f the experimental eUITent loop, a speeial transformer is used . To eope with the large reaetive power required to energise long earth fault eurren t loops, a eapaeitive reaetor bank ean be switehed in parallel to the output of the transforrner, thus a llowing the eompensation of a maximum reaetive power of 550 kVAr. The eleetronie frequeney eonverter and the transformer are designed to injeet eUITents at frequeneies from 12 Hz to 500 Hz, enabling for example the study of the frequeney depende nee of the impedanee to earth, ar working elose to railway frequeney (Le. 16.7 Hz in Switzerland).

3 Problems with measuremeots Besides interferenee problems, there are some fundamental problems with earthing measurements whieh ean eause signifieant uneertainty in the interpretation of the measured quantities. Additionally, there are some inherent limitations of the reprodueibility of sueh measurements. The problems responsib le for these ineonvenienees are listed and, where applieable, eommented in the following . Some of these open prablems ean be overeome by using a standardi sed measurin g proeedure. Speeial "speets of these rneasurement proeedures should be exarnined \Vith experimel1ts an eanhing system s:

I. The 5pecific re5i510l1Ce 0/ Ihe 50il is nat eonstant in time mainly beeause of different humidity and temperature. It is subjeet to a variation of roughly one order of magni tude .

2. Due to the compressiol1 0/ Ihe 50il, the resistanee to earth of newl )' la id eanh eondue!Ors is slowly redueed, a proeess whieh ean take many months .

J. Eanh fault eurre nts do nat produee a spherieal DC eonduetion field arou nd the eanhing system beeause of the il1holllogel1eoll5 soi! COl1dllClivilY. For geologieal reasons, the soi l eondueti vity deereases with depth, whieh leads to surfaee- bound eanh eurrents. Negleet in g AC effeet s, the eurrent dens ity and the resi stive eleetrie field of a single eanh eleetrode deerease with the square of the distanee from the eleetrode for small distanees, and almosl linearly for larger dislanees. for simplieily, we ass ume an earth eleetrode with a boundary of a half- sphere for the following eon­siderations. The earthing voltage UE is the integral of the resistive e leetrie field strength EE (r) taken from the boundary ofthe earthing system (r = ro) to an infinitely large distanee .

u = [E (r)dr= [Eo·(!...) -kdr E ro E rO r

o l <k < 2 (I )

EE resistive eleetrie field strength around the earthing syslem [V/m] r distanee from eentre ofearthing system [m] ro radius ofeanhing system [m] Eo res istive eleetrie field strength on the boundary ofthe eanhing system [V/m] k exponent of the deereasing potential funeti on depending a n geologieal faetors

The earthing voltage UE and the impedanee to earth ZE = UE / lE of the installation depends upon the distanee lo whieh the inlegral (I) is evaluated (in prac ti se lhis ean never be infinity). In the

3.2.3

hmlllng casc 01' a currclll propagalloll on Ihc 5urfacc of Ihc 50 11 (k :-:: I l. {he Impeuancc to crln h

\Vould evell become infinilely large (see also [6[. ehapler D. elause f)

4 rhe problern oF Ihe uneer1am definilion oF Ihe impedanee lo car1h deseribcd above (po"" .1) beeomes even more eompliealed iF we eonsider Ihe Fael Iha I car1h Faull eurrelllS are i\e eurrelllS which are subjeel 10 eurrenl displaeemenl (skin effeel) in Ihe soíl (see figure 5). Ahernaling eurrenls in Ihe soil do noi spread oul very far. bUllhey slay as elose lo Ihe loop oF Ihe fault Currenl as possible. Ae currenls do. IhereFore. noI propagale in Ihe shape of Ihe elee!rie field oF a slalie dipole (figure 5. IOp). bUI in a straighl eorridor oF a widlh oF some 100 m in Ihe direelion oF Ihe line Feeding Ihe Faull eurrenl (Figure 5. bOllom) . The resull is a slrongly asymmelrical earlh currel/I dislribuliOll around Ihe ear1hing syslem. partieularly sinee Ihe experimenlal Fault eUITenl is Fed only From one side .

The penelralion deplh 5 in Ihe soil oF Ihe magnelie field ar Ihe eUITenl yields quanlilalive inForma· lian an Ihe propagalion oF earth eUITenls.

õ p w

161 = J'X,l1o penelralion deplh oF earth eUITenl [m J speeifie resislanee of soil [!lmJ angular frequeney ofearth cUITenl [SIJ magnelic perrneability ofvacuum [471.JO·7 VslAmJ

(2)

Penetration deplhs õ between 150 m and 1500 m result fram typieal values of Ihe specific resis· lanee p of the soil ranging fram 10!lm lo 1000!lm at 50 Hz. Fram Ihese statemenlS il ean be coneluded thal Ihe measured impedanee to earth of an installation does nat anly depend upon Ihe evalualion dislanee (aceording lo painl 3), bul al so upan the direction of Ihe measured vollage profile wilh respeello Ihe feeding line. and an the frequeney. Therefore, partieu[arly [arge subsla· lions wilh dimensions in Ihe order af magnilude of Ihe penelralion deplh Õ. i.e. some 100 111 .

possess an earth eurrenl dis!ribulion whieh s!rong[y depends an [he direelion of [he fault eurrelll feeding [ine . Furthermore. also Ihe measured loueh and slep vollages s!rongly depend an Ihe [ine ehosen lo feed Ihe earth fau[1 eUITenl. This fae[ eomplieales Ihe evalualion aflhe earthing silualions of [arge insla[[alions if only one configuration af feeding [ine has been exam ined.

5. Often [here is na elear eanfinemenl af Ihe earthing syslem beeause Ihe inslal[a[ion eansisls aF severa[ remo[e parts (e .g. a terminal pa [e af a averhead [ine) whieh are eannee[ed by earth eondue· [ars (aver and under ground). For Ihis reasan, Ihe impedanee [o earth is arten na[ e[ear[)' defined . Fina[[y, beeause af [he abave menlianed remale parts af an insla[[a[ian, Ihe resu [l s af [he measuremenls af [aueh and slep va ltage depend slrong[y an Ihe direclioll al Ihe lallll CIII'I'C/lI leedillg lille.

The abave menlianed aspeels elearly demanslrale Ihal Ihe real earthing situalian af any insla[[a[ion ean an[y be simu[aled by injeeling experimenlal eurrenlS via all earth fau[[ eurrenl feeding [ines. For praetiea[ reasans, Ihis kind af simu[atian eannal be app[ied to a standard examinalian. There· fore, lo aehieve reprodueib[e resu[ls, il is impartanlla use a slandardised procedure for [he exami· nalion of earthing syslems af [arger inslallalions . Today. Ihough, Ihere is a laek of sueh guidelines or recommendations.

4 Elimination of intcrfcrcncc

High vollage subslalions and [arge power generalion p[anls usually remain lO operalion during earthing syslem measuremenls. Therefare. by ahmic ar induelive eoupling, power frequeney interference ean be eaup[ed inlo Ihe experimen[a[ eUITenl [oop ar any measuring [aap. Thi s inlerferenee may lead lo subslanlia[ eITors in Ihe measuremenl if il is nat e[iminaled properly [4, 51 . Besides mains frequeney inlerferenee (50 Hz, 60 Hz, 16.7 Hz) a[so iIS harrnonies and De eurren[s have lo be eonsidered . The fallowing [eehniques are wide[y used lo suppress inlerferenee (see a[so ([ 1,[2 J.[31J

3.2.4

• hcat rncthod

• on/off melhod • rolarily reversal melhod • vcelor voltmctcr, synehronoU5 rcctificr , loek-in amplificr

• ehoiee of a faull eurrenl frequeney differenl from mains frequeney. li'l!enng of inlerferenee signals

A summary of Ihe advanlages and disadvanlages of Ihe differenl melhods for inlerferenee eliminalion is given in Table I .

The beal melhod uses an experimenlally injeeled earth faull eurrenl wilh a frequeney slighlly differenl from Ihe operaling frequeney of Ihe inslallalion (devialion smaller Ihan I Hz). A bealing is produeed by superposilion of Ihe mains frequeney inlerferenee vollage U, and Ihe measuring vollage Um , whieh leads 10 an oseillalion of the read-out of the voltmeter belWeen a maximum value Umax and a minimum value Um ;n' The measuring vol!age Um ean be ealculaled aeeording to Figure 6 . It is evident Ihat the aeeuraey of this method depends on the skill of the person who eaITies out the measuremenl and on the time eonstant of the voltmeter.

The onloff method employs a main 5 frequeney fault eUITent whieh is switehed on and off periodieally. The melhod is ealled polarity reversal method if the IWO phases of the eUITenl souree are reversed before eaeh on-eycle. AII measurements are eaITied out in Ihe on- and off-state (U,) of the eUITenl souree; in Ihe ease of the polarity reversal method, the measurement is perforrned at both polarities (Ul al positive and U] at negative polarity) and in Ihe off-slate (U,). 11 has 10 be menlioned Ihal Ihe polarity reversal method only delivers reasonable results if Ihe experimenlal eUITenl is absolutely in phase wilh the operaling frequeney of Ihe inslallalion. Moreover, Ihe inlerferenee has lo remain eonstant during the three readings at one partieular loealion.

Whereas Ihe on/off melhod gives only an upper limil for Ihe measuremenl error, Ihe polarily reversal melhod ean yield an unambiguous relalion between the Ihree measllred values and Ihe vollage U", eaused by Ihe injeeled eUITenl (see Figure 7) . Reeenlly, a eomputer supror1ed on/off method whieh allo\Vs an effieienl suppression of inlerferenee has been presenled [5J.

In the lasl few years, Ihe aulhors have galhered a 101 of experienee with all of Ihe melhods menlioned exeepl for the polarity reversal melhod . The mos! effeetive elimina!ion of inlerferenee was aehieved \Vilh Ihe use of a fault eUITent wilh frequeney different lo mains frequeney. In some eases Ihis melhod is used in eombinalion wilh synehronous reeliliers (i.e . loek-in amplilier), Ihus resulling in a very high signal-Io-noi se ralio .

The experimemal ear1h faull eurrenl is eilher produeed with a diesel generalOr or a slalie frequene)' eonvCr1er. For subslalions operaling al 50 Hz, a frequeney of Ihe injeeled eurrenl of 70 Hz is usuall y ehosen. For inslallalions of Ihe Swiss rail\Vays (SBB), \Vhieh operale al 16.7 Hz. measuremenls are generally performed between 12 Hz and 24 Hz. This devialion from Ihe operaling frequeney of Ihe inslallalion under les I allows a very simple and effeelive separalion of Ihe experimenlally injeeled eurrents from earth eUITenls produeed by the operalion of Ihe plan!. The measuremenls of toueh and slep vollages are eaITied oul wilh a speeial voltmeler ineluding highly seleelive nOleh lilters for 50 Hz and 16.7 Hz.

For the measurement of earth eUITenl distribulions, loek-in amplifiers and synehronous reelifiers have been used for Ihe very sensilive measuremenl of amplitude and phase of par1ial ear1h eUITenls. This sophisliealed inslrumentalion enables the aeeurale assessmenl of large high vollage inslallalions al only moderale injeeted eurrenlS (Iens of amperes). However, Ihe Iransmission of a referenee signal from the eurrenl souree lo Ihe plaee of measuremenl is a prerequisile for these leehniques.

3.2.5

Mellu,,1 Freqllellcy MeaslIrelllellt A dvalllages Disadvfllltages

1 13c<u mcthod Dcviation [rom Mcasuring time a Mcasurcmcnt dose 10 mains Evaluation nOI always mains frcqucncy fcw scconds. with- frcqucnC)' unambiguous < I Hz out filtcrs

2 On/orr mcthod Arbitrary 2 mca'iurcmcnls, Simplicit)' No min imisation. but without filtcrs anly cSlimation or crror

J. Palarily rcvcr- Mains frcqucncy 3 mcasurcmcnts, Unambiguous dctcrmination or No climination orhar-sai mcthod without fillcrs mcasu ring voltagcs. men- monics. 3 mcasurcmcnts

suremcnl at mains frcqucnc)' nccdcd

4 Vcclor \10 h- Arbitrary One mcasurcmcnt Supprcssion ar all kinds ar Rcqucsts refere nee mete r laek-in with dedicatcd inlcrfcrcncc possiblc signal [rom cxp. cu rrenl amplificr instrument sourcc

5. Filtcring ar Dcviat ion from Onc mcasuremenL Suppression of al1 kinds or Measuring frequcncy intcrfercncc mains frcqucncy with filtcrs intcrfcrcncc possiblc nOl identical to mains

>10 Hz frcqucncy

TABLE I Compilation of methods for the suppression of interference and their advantages and disadvantages.

5 Detcrmination of vo1tage profiles and impedances to earth

Earthing voltage profiles are a valuable too I for the further assessment of earthing systems. An earthing voltage profile is determined by measuring potential differenees between the earthing system of the plant (referenee) and metal pike eleetrodes whieh are inserted into the soil at defined positions (e.g. a l I m, 5 m, 10 m, 100 m distanee from the boundary ofthe plant) in a given direetion.

Figure 8 sho\Vs the result of sueh an earthing voltage profile measurement at different frequeneies of the injeeting current. The frequeney-dependenee ofthe potential profiles is due to displaeemcnl effeets ofthe earthing currents (see also seetion 3, paragraph 4).

From the shape of the voltage profile, the 'hot zone' around the installation and the effeetiveness of surfaee potential eontrol measures ean be verified. Partieularly the vollage gradients are c1early visible in this representation. The exisling step voltage ean be taken as voltage differenee from the measured vo ltage profile.

The effective impedanee to neutral earth ofan installation ean be determined by evaluating the voltage profile along a suffieiently long distanee. The value of the impedanee to earth ZE is defined as the ratio of the maximul11 (asym ptotie) va lue UE of the voltage profile to the earthing eurrent lE whieh produees this voltage:

(3)

An important reason for the measurement of voltage profiles is the question of the magnitude of the earthing voltage Ur;. If UE is sma ller than the maximum permissible toueh voltage, the measurements ean be stopped, i.e. no toueh and step voltages have to be measured. Field experienee eonfirms that the vo ltage profile strongly depends on the direelion of the profile relative to the earth fault eurrent feeding line, as has been stated in a previous ehapter. Reprodueible results may be obtained if an angle of about 90° between the direetion of the profile and the feeding line is maintained (Figure 5b) and if no other eondueting struetures are in the vieinity of the profile.

If all of these eonditions are fulfilled, Ihe voltage along the profilc reaehes saturation eaused by eurrent displaeement (skin effeet) after a few hundred meters. Therefore, a elear and meaningful value for the 'apparent' earthing voltage UE ean be determined. The measurement of at least two profiles is always reeommended sinee this allows Ihe reprodueibility ofthe results to be eheeked .

3.2.6

6 Measurement of lourh, slep and differential voltages

Toueh vohages are measured bel\Veen Ihe poinl of eonlael and Ihe (Ioeal) eanh in a drslanee of I m of Ihe melal objeel (Figure I) . To eonl.el Ihe e.nh .• pike eleelrode (senling as referenee potenl,al) " driven inlo Ihe soil 10 a deplh of. few eenlimelers . In een.in eases. differenli.1 voltages bel\Vecn I\V(I objeels are measured as \Vell iflhey are nexl 10 eaeh olher .nd iflhey are noI eleelrieally eonneeled

To perform eanhing measuremenlS. a dediealed seleelive vollmeler was developed (design by FKH) This inslrumenl includes seleelively swilehable band slop tillers for 50 Hz and 16.7 Hz and allo\V, direel readings lo be laken of vollages from 10 mV 10 1000 V. The fihers suppress inlerferenee from 50 Hz or 16.7 Hz by abou I 40 dB. enabling Ihe measuremenl of signals in Ihe millivoll range. T\Vo readings are usually laken for eaeh measuremenl posilion : in Ihe firsl measuremen! Ihe voll.ge is measured direelly over Ihe high inpuI impedanee of Ihe aetive filler; in Ihe seeond measuremenl Ihe input of the tiher is loaded by 2 kO. The firsl measurement yields the voltage aetually presenl al a eenain position (i.e. the prospeelive loueh voltage [I D, whereas Ihe seeond one provides the voltage whieh would appear iflhe measuring posilion was loaded wilh Ihe impedanee oflhe human body. The voltage values taken with the 2 kO-load are usually taken for funher analysis.

The effeetiveness of the eanhing syslem and toueh and slep vollages appearing in ease of an eanh faull have lo be assessed and summarised in a repon. Typieally, Ihe toueh and step vohages are represented in Ihe form of bar graphs as shown in Figure 9 and in Ihe map in Figure 10. The values of the measured voltages are linearly sealed to an eanh fault eUITent of I kA or to the maximum eanh fault CUITenl of the installation.

Countermeasures have lo be taken if inadmissible toueh or slep voltages have been recorded. The following eounlermeasures are often taken lo prevenl high loueh or step voltages or to reduee their danger for people:

• laying an addilional surfaee layer (asphall, erushed roek) • surfaee potential eonlrol wilh additional eanh eleetrodes • eoating the metal strueture wilh an insulating painl • putting up ofwaming signs • redueing proteelion Irip time.

7 Determination of earthing eurrenl distribution

The determination oflhe eanhing eurrent dislribulion provides answers 10 Ihe following quesli"ns :

I) Are.1I oflhe eanhing eonduelOrs intael and of low resist.nee?

2) Are there power or signal eables whieh are loaded by exeessive eanhing eUITenls (shealh eurrenls ) in ease of an eanh fault. whieh may lead lo penurbalions, ovenlollages or even Ihermal overload of a eable?

J) Are Ihere unwanted eonneetions belween insulated conduelors an zero potential and eanh (open cable shealhs, separaled eanh eleetrodes)?

A syslematieal determination of the eUITen! distribution of all eonduelors whieh are in loueh wilh Ihe earthing system oflhe inslallalion ar whieh are used for iIs earthing, respeelively, is reeommended for Ihe assessment of Ihe effeetiveness of earthing syslems. Table 2 lisls Ihe relevant types of eonduetors .

If all of Ihe eUITenl shares in Table 2 are registered in magnitude and phase and if Ihere are no olher eonneetions between Ihe surrounding earth and Ihe earthing syslem of the installation. Ihe veelorial sum of all regislered currents musl be equal lo Ihe injected CUITen!. By means of this measuremenl, the distribulion of Ihe earth faull eUITent ean be assessed. Generally, a subslanlial fraelion of Ihe earthing CUITenl flows direetly from Ihe earthing grid inlo Ihe soil. This fraetion eannOI be measured and appears in Ihe 'balanee sheel' as the non -assessed remaini/lg curre/ll . An example of an earthing eUITent 'balanee shee!' is shown in Figurc I I.

3.2.7

• eonneelions lo eanhing eleelrodes and eanhed slruelures

• eanh wircs of overhead lines

• earthed shealhs of power eables

• low vollage eables

• control and signal cables wilh eanhed shealhs

• lelephone eables

• eondueting water pipes and gas eonduets

TABLE 2 Lisl of types of eonductors to be considered for Ihe evaluation of Ihe eanhing CUITen! distribution.

The measurements of earthing CUITent shares are performed by means of prong-type ampere meters, and for eonduetors of larger diameter (e.g. high voltage cables, pipes and high voltage poles with diameters up to I m) with dedicated Rogowski eoils (see Figure 12). To avoid interferenee, the voltage aeross the burden of the current transformers has to be transmitted by means of twisted and shielded eables (twinax-eable). As in the case oftoueh and step voltages measurements, band stop filters ean be used to suppress 50 Hz and 16.7 Hz interferenee (see seetion 6).

The deterrnination of the phase of eanhing eurrent shares in earthing eonduetors ar eable sheaths reguests a referenee signal which has to be transmitted from the fault eurrent generator to the plaee of measurement. For distanees up to about 100 m, this referenee signal is transm itted by means of a twinax eable to the plaee of measurement, where the phase shift between this referenee and Ihe eanhing eUITent share is determined. For Ihe lransmission of referenee signals over larger distances (up lO I km), signal lransmission by radio was developed. A shon synehronisalion impulse is lransl11itled by radio (FM 434 MHz) al eaeh deteeted zero erossing of the injeeled eurrent, hence allowing the determination ofthe phase shift ofthe measured eurrenl. Assuming thal the phase shift of Ihe referenee ehannel has been balaneed prior to the measurement, the phase angle ean be determined with an uneertainty of ±5°.

The earthing eurrent shares are divided into a real and an imaginary part, permitting lheir veetorial sUl11l11ation (Figure II l. The eOl11ponenl of Ihe eurrent in phase with the injeeled eurrel1l is ealled the real parl, whereas Ihe componenl shifted by 90° is the imaginary parto

7. I "Indirect" measurement ar carthing curren!s in ear!h wircs of overhead lines

Earlhing eurrent shares in earth wires of overhead lines ean be assessed by measuring the l11agnetie induclion B due to lhis eurrenl. This indirecl assessmenl has Ihe advanlage lhat eurrents in hardly aeeessible earth wires ean be measured on ground level by means of a loop antenna. Again, mains fregueney interference ean be separated from the injeeted current (e.g. 70 Hz) by mean s of band SlOp filters.

The prineiple of the 'indiree!' earthing eurrent measurement is shown in Figure 13 . The earth wire eurrenl l]' generates a magnetie induetion B, whieh induees the voltage U;nd in the loop antenna. This indueed voltage is amplified and inlegrated in time, Ihe resulting voltage U""as is proportional to Ihe eurrent Oowing in the earth wire and inversely proportional to the dis!anee d to the eurrent path.

The integrator ean be ealibrated in sueh a way that the eurrent Irean be caIculated aeeording to

(4 )

Even very smalI eurrents in the milIiampere range ean be l11easured with this l11ethod if synchronous reetifiers are used, no matter whelher Ihe overhead line is in operation or no!.

3.2.8

8 Conclusion

Thc appliealion of supplemenlary experimenlal methods (primarily by measuring Ihe eurrenl dislri­bUlion) resulls in a mlleh better underslanding of Ihe earthing syslem for Ihe ease of an earth faull. The knowlcdge of Ihe eurrenl dislribulion among all the eonduelors leaving a high vollage subslalion or power planl gives informalion on possibly protraeted potentials. It helps to indieate plaees where high toueh voltages are to be expeeted or supplies explanations for high potential gradients or toueh vollages, respeetively. A systematie assessment of all earthing eurrents al so helps to deteet eonduetors that earry high eurrent shares and might possibly be overloaded in ease of an earth faull.

Attention has to be paid to measuring errors on aeeount of interferenee with the operational eurrents and voltages in the examined installation whieh may be indueed induetively or resistively. It has been demonstrated that the ehoiee of a fault eurrent and measuring frequeney s lightly different from the normal operational frequeney of the power network offers refined possibilities to eliminate disturbing interferenee. Simple noteh filters may manage this task sueeessfully. A more eomplieated but eertainly effeetive method to measure small signals in a noisy environment is the applieation of a loek-in ampli­fier (synehronous reetifier). With this method, the badly eonditioned measuring signals are eompared eleetronieally with a referenee signal of the injeeted fault eurrenl.

Due to inherent experimental limitations it is obvious that the examination of an earthing system does not present preeision data sinee the existing measuring methods are not strietly defined and sinee the results depend on many details of the experimental arrangement (i.e . direetion and size of the fault eurrent loop). Thus it is not permissible tojudge the earthing system ofa HV-substation with kA-fault eurrents as "suffieient" or "insuffieient" on the basis of a few toueh voltage values measured. A pivotal problem is that an earth fault may often not be simulated realistieally enough. The earthing eurrent distribution always depends strongly on the ehoice of the power line where the experimental eurrent is fed in. The eurrent displaeement by its own magnetie field is one of the most important effeets that eauses an asymmetrieal eurrent distribution and eomplieates the analysis of the measured potential raI se.

The melhods for experimental examinations have to be established in sueh a way Ihat all results are reprodueible and thal as little seope as possible is left for the interpretation of the measurements. In this respeet, preeise reeommendations for the execution of earthing measurements are highly desirable.

Annotation The full eontenl oflhis eontribution has already been Pllblished in German in Ihe Bulletin ofthc Swiss Eleelroleehnieal Assoc ialion [7J . The allthors address spec ial Ihanks lO Dr. Thomas Heizmann, FKH, for comprehensive help in doing Ihe lranslation ofthis article .

9 References

[I J European Standard Draft prEN 50179 "Power inslallations exeeeding I kV AC" European Committee for Electroleehnical Slandardization CENELEC, December 1996.

[2J " Erdungen in Slarkslromnetzen" ; Vereinigung Deutseher Eleklri z itiitswerke m.b.H. - VDEW, Frankfurt am Main; J. Auflage, 1992.

[3J A.P.S. Meliopoulos: "Power System Grounding and Transients : An Introduction", Marcel Dekker, New York, 1988.

[4J F. Sehwab: " Erdungsmessungen in ausgedehnten Anlagen"; Bullelin SEVNSE Vo!. 71, No 4, (1980), pp. 174.

[5] R. Hoffmann: "Neues Messverfahren zur Eliminierung von Fremd- und Stõrspannungen bei Beeinflussungs- und Erdungsmessungen"; Elektrizitiitswirtschaft, Vol 91, No 22 (1992) pp. 1455 .

[6] W. Koch: "Erdungen in Weehselstromanlagen über I kV, Bereehnung und Ausftihrung"; Springer Verlag, Zweite Auflage, Berlin, 1955.

[7] R. Brãunlich: "Die messtechnisehe Überprüfung von grossen Erdungsanlagen"; Bulletin SEVNSE, Vo!. 87, No 23 (1995), pp . 31.

3.2.9

S'ome nUlionu/ slal1darcújo1' earthil/í!. direcliml.\'

181 ANSIII EEE Std. 80-1986: "IEEE-Guide for Safety in AC Substation Grounding" .

191 CP.10 13 1065: "British Standard Code for Praetiee, Earth ing" .

[101 IEEE Std. 81-1983: "IEEE-Guide for Measuring Earth Resistivity, Ground Impedanee, and Earth Surfaee Potentia ls

of a Ground System".

[111 DlNIVDEOI4117.76: "VDE-Bestimmungen fur Erdungen in Wechselstromanlagen rur Nennspannungen über I kV" .

[1 21 Regeln des SEV 3569-1,2,3 : 1985 :

2: ::J_ .. '" S Õ >

.<:: u :J B .. > ~ " c. " o ~

a.

"Erden als Sehutzmassnahme in elektrisehen Starkstromanlagen, Teile I bis 3" .

1000

100

10

0.01

- - - r - .... . • t· .. •

- :-. I

- ;.--

I , _ .

--CENELEC prEN 50179, 1996

- - Swiss standards: SEV 3569-1 .1985

0.1

~ . • i . "

1

Maximum duration 01 touch voltage t F [s]

Figurc I Permissible touch voltages U71 , for limitcd fault curfen! now duration.

3.2. 10

10

Impedance to earth ---

lE

.. . , . : : i ; ':·

o 411-----Touch voltage

Step voltage

~ - -rJ~1~ ---: _. - - - - --_ .... -- - - _ .....

, \ I

Figure 2 General principle ofearthing measurements. An earthing fault current 'E is injected into the earthing system. This eUITent flows through the soit and eauses a potential rise UE. in the vieinity ar the earthing system. Touch and step voltages are generated in the 'hot zone' around the instaIlation.

Return current by earth wire

IL - ..

Dlsconnecting swilch r, Dlsconnectmg swilch

--~~...-t-----~/f----~--~~~~~------~~~~~--~ ~~----4 ~~ __ _

Expenmenlal current I E

Adaptation Iransformer

Overhead line used for current injection

Return current by s011

~ - -- -~-

Earthing system 01 installation under tes~

Ear1hmg SWllch

Earthing system 01 second installation: auxiliary earth

Figurc 3 Fonnation of earth fault current loop by means af two substations and a connecting overhead linc. The sum of the return current in the soi! '1/ and the return currents in the earth wirc l,. rcsults in the experimentaIly injected earth fauIt current l,:".

3.2.11

Gas.: iVsl ' iVml /

/ Um lI; \ f-_U_m---l:=-:-1'~U ~ . ~ '1 U

m

'" ' 1 umu

~ ~ -U +U m m

[ => _ _ u_"_=_Fu_(_ur. (I: ' i1'. u: U, ,u] : Measured vofrages U.: Interference

t-Um ,-u,": VoltiJges genera led by experimental current

3.2.12

Figurc -t Currcnl source con slsling 01" • frequeney·convener (50 kVA) lhc housing includcs auxdiary installat ions and cOf1trol circuits .

a)

\ ,

In.llIlIlIllon 1 Source

bl

AC

Figurc 5 Lines ar current in a resistivc eleetrie field belWeen 1W0 eanhing systems (qualitative representation) for both eases: AC top; DC bottom. An eanh fault has been sel up In

installation 2, which is fcd ovcr a transmission !ine from installation I. Notc the coneentration af current Jincs in Ihe ease of AC an Ihe side of Ihe line depanure.

Figure 6 Detennination or Ihe measuring voltage with the beat method . Thc measuring voltagc Uit! is detcnnined by eliminating the interrercnee with mains frequcney for Ihe ease: that the measuring voltage is larger Ihan (he inlerferenee (Iefl), and for Ihe ease Ihal Ihe measuring voltagc is smallcr than Ihe interfcrcnce (righl).

Figurc 7 Determination ar {he measuring voltagc with the polarity rcversal method. Vector diagram explaining the elimination ar mains frequeney inlerferenee (derivalion, see [5J).

60

-c: Q)

~ 50 ::J U

~ ::J

~ .c: t ro Q)

ro :;:: c: Q) -o a.

40

30

20

10

o O 50 100

-~- - <> .. -

0- _ 0- _ 0- _ 0

-0 - _ _ o _ - .. o- .... o .... - o .. 0 __ 000 - .. - 0 --

- o - 12 Hz

-0 -35 Hz

-0-70 Hz

150 200 250 300

Distance from substation [m]

Figure 8 Voltage profiles of a 132 kV-substation measured at three different frequeneies of the injeeting fauh eUITent. The impedanees to earth were estimated to 27 mO at 12 Hz, 41 mO at 35 Hz and 55 mO at 70 Hz.

Pos ition No . on map

~ 35 3 4 56 7 8 91011 1213 t41516 1718 192021222324252627282930313233 34 3536

E " ~ :; 30 u 111

:J

~ 25 I- Legallimit for touch and s tep voltages for protection trip time of 0.1 s - , 'õ (Ug = 700 V) scale d to an earth fa ult current of 1 kA

~ 20 ~

2 " 15 c)

!l Õ 10 > a. " iií (; 5

" U :J

O o

- !i- - - - - - - - - - - - - - - - - - ' -I ,

~-1- ; { ! ! . ' j . \ ,

i I ~ :f-- ( ~ 'f-- : I , ; ~ , , , - , ,

\ i ; ,

; , , }

, [ , · , i J • I

, , i l

, ; ,.

~ \ ~ I , ,

~ · i , i , j ,

~ , I , ,

:1 ~ \ ,

m j {

, ~

, , ~ ~

; , ; ffi '" f;I n .' ;

I l , i i ! , i , [ "' : r ; 111

, , , , , ; n , , , , · ... ~

~ ~ • · E • o o o o E ~ o o o • o ~ ~ ~ ~ E e e e

ª i:' ~ ~ • • • • S E .~ S ~ .S S <1

< , ~

e

~ ., E E ~ ~ ~

g ~ o

~ ~ .~ ~ B E e • 'E • E E li. E E " li. • • J~ " § ~ ~ < ~ • o • n ~ 9 ~ ~ ~ ~ ~ • < < o

~ ~ S

~ n li n n n

j j ~ ; j e ~ ~ ~ " . r õ õ li ~ li li " ~ E ~ ~

E E E ~ • li o , i! , , E E E g g g • • • • • · o o o • , u e o e e Õ Õ Õ u u u u o • • ~ Õ e e e g -; u 5 5 • g g .g õ Õ n n . u ~ e o B B 5 u

~ li e • • ª ª ~ ~ e • e • e • Õ Õ Õ • ~ ~ ;; ~ " ~ w 'E w a: ~

5 5 • = = S S • ; n • • õ õ õ Õ e ~ m ~ • 5 • i e u u e e li u Des ignation of object • • ~ ~ ~ ~

Figure 9 Graphic representation of measured touch and step voltages in the surroundings of a high vohage substation (for measured positions, see Figure 10).

3.2. 13

" ." .,0

.. 0°:' " ,

- - --: , ', '. J'.'. ",,'-, "" '"

.--_ ..

Figur. 10 Map or high vol!age substation (400 kV·GIS ins ide building). The loeations where toueh vo l!ages were measured are marked and labe lled with a position number eorresponding to Figure 9.

en ., '­:l 1:: ni C. .,

1J ., .o ni u '-O en '-O ~

U :l 1J e o u cl e .c 1:: ni

W

-400

... Remaini g curren ~ Ea<lh;' ndu O"

.... Coonftcl,on po .... er 51oll!0f\ . ''''' cl'>9B .1r l

..... Connec! l)n poWCf 51011 00n • ... ,tcngear 2

r-Eanh'l\ÇI o .)ltU "

Wa ler s uool ''" ing cooduClOt dlre loon compe l'lU[O -~ Wal!!r P'PI! en • .... 9 bu.1tlorog

Siona l cahl

Coa. I cat>le

Cable o w. r ca lc h menl

SIOnal cabll! lO .., ler calCl"omenl >L l ow vo lta o. deoartures

Low ... oll a e Il'ISl allahcn 01 w al(~r e [!;.hmonl

Sewage pump

Sl rcOI hgnlong

OeQªDure ~O kV ~ u bs talion

~ Oeparlvfe .... ell

Linc fQrmin g urrcnt 1000. O kV

Inject d curren Rl!lurn tuUenl ... sncaLn

1kA 1./ ·300 ·200 .1 00 O 100 200 300 400

Current shares [A)

_ R.(IIIAI p.r I kA = Im(II IAI p. r IkA

Figurc 11 Example or a 'balance sheet'. Vcclo rial summ ation is rcquired bccause a r substan tial phasc shins belween Ihe d iffercnl shares of Ihe carthing current.

) .2. 14

Figure 12 Measurement or a current share in a pole or an overhead line (termination or a high voltage cable).

Magnetie Induetion

t d

~ Surl,c. o, ,,'eoo, '\ A

Loop antenna ' '::::::::;:::;:::1/

Induced voltage

B Earthmg current to be measured (f:150 Hz)

Earth wire

. __ ._----,

f] = __ 2[1 Jf ~Id I

Integrator

r-·------------{

I ~",,,a, = ko . f uu,,, . dI

j Referenee slgnal from expenmental current souree

'-- .~

I

Phase sensitive voltmeter or

digital oscilloscope

Figure IJ "lndircCI" dctcrmination ar t:arthmg currcnt sharcs in ground wircs ar a vt:rht:,td IIIlt:s

32 . 15