This is the ccepted mnscript, nd which hs been ccepted by IEEE for pbliction. IEEE. Personl se of this mteril is permitted. Permission from IEEE mst be obtined for ll other ses, in ny crrent or ftre medi, inclding reprinting/repblishing this mteril for dvertising or promotionl prposes, creting new collective works, for resle or redistribtion to servers or lists, or rese of ny copyrighted component of this work in other works. The fll reference is: Incresed Energy in Stble Dry-bnd Arcs De to Length Compression X. Zhng, S. M. Rowlnd nd V. Terzij IEEE Trnsctions on Dielectrics nd Electricl Insltion, Vol. 7, Isse, pp. 7-8, April Digitl Object Identifier:.9/TDEI..8
Incresed Energy in Stble Dry-bnd Arcs De to Length Compression X. Zhng, S. M. Rowlnd nd V. Terzij School of Electricl nd Electronic Engineering The University of Mnchester PO Box 88, Sckville Street, Mnchester, M6 QD, United Kingdom ABSTRACT The occrrence of dry-bnd rcs on otdoor composite insltors cn degrde the polymeric mterils srfce nd ltimtely my led to insltor filre. The degrdtion processes re generlly considered s ging effects occrring over long periods of time, from yers to decdes. In this pper, it is shown tht if stble drybnd rc is physiclly compressed in length by externl forces, sch s electrolyte deformtion de to wind or grvity, the rcing ctivities will become more severe. This in trn, my ccelerte the degrdtion into short time-frme hzrd. A series of experiments re crried ot to investigte the electricl chrcteristics of the rcs s they become compressed. In this cse experiments re performed on silicone rod insltors t controlled ngles to the horizontl. Rpid ging is observed fter sch events. Mesrements of rcing period, pek crrent, nd rc resistnce dring the rcing compression process re nlyzed. Bsed on the experiments, Doble Sinsoidl Model is developed to simlte the crrent-voltge chrcteristics of drybnd rcing dring its compression. Both experiment nd simltion show tht rc power, rc energy nd corresponding energy density will drmticlly increse if rc compression occrs, which my led to more rpid nd serios dmge on composite insltor srfces thn is experienced otherwise. It is sggested tht ggressive erosion events my occr in short periods of time within extended test regimes or entire service histories. Index Terms Composite insltor, silicone rbber, ging, NCI. dry-bnd rc compression, model, rc energy. INTRODUCTION IN both service nd experimentl conditions, electricl dischrges sch s dry-bnd rcings my develop on insltor srfces de to electric fields, the presence of moistre, low srfce hydrophobicity nd polltion []. Dry-bnd rcing is one mechnism responsible for the chemicl chnges to the srfces of polymer insltors, which my led to hydrophobicity loss, erosion of the srfce, nd even penetrtion reveling the fibreglss core nd ltimtely csing mechnicl filre of the insltor [-]. Dry-bnd rc crrents re normlly limited to the order of millimps de to the resistnce presented by the rest of the insltor srfces. If crrents exceed ma, flshover becomes possibility. Dry-bnd rc resistnt mterils hve been developed over the yers which re ble to limit reslting degrdtion nd erosion [6]. Therefore, the deteriortion from dischrges inclding dry-bnd rcing is generlly considered s long-term ging effect on good qlity composite insltors, nd these mterils cn withstnd dry-bnd rcs for mny hndreds or even thosnds of Mnscript received on My 9, in finl form 7 Jly 9. hors in lbortory slt-fog tests. As reslt of reserch on ccelerted dmge on ADSS (ll dielectric self-spporting) cbles designed for high voltge environments [7-8], novel view hs been proposed tht ccelerted degrdtion on insltor srfces cold be possible if dry-bnd rc is physiclly compressed in length. Sch n rc-compression event cn reslt from the movement of moistre t the edge of stble dry-bnd over which n rc is strck. The rc contribtes to the likelihood of moistre movement becse it redces the srfce tension of the wter t the rc roots. In this pper it is shown tht dry-bnd rc compression reslts in more energetic rcs nd sch compressed rcs cn degrde insltor srfces mch qicker thn the norml grdl degrdtion expected. As reslt it is proposed tht locl deep erosion ptterns seen fter extended tests or periods of service cn be de to rre rc-compression events rther thn grdl erosion. A series of investigtions for dry-bnd rcing chrcteristics re reported in [9-]. Those stdies were bsed on horizontl smples where the dry-bnd rcs cold grow to their ntrl fll length. Inclined plne testing with contminnt flowing down slope to n rcing loction [] ntrlly compresses
the rc nd chnges the rcing properties in comprison to the horizontl sittion. This is nderstood to be more ggressive test. The proposed mechnism in this pper lies somewhere between these two sittions. Dring the proposed mechnism [8], therml nd moistre blnces re reched llowing formtion of stble dry-bnd rc with repetble electricl behvior in every power cycle (Figre ). When the smple is inclined (Figre b), the pper wter film nd lower wter film, whose edges re t the dry-bnd bondries, will tend to move down the insltor core de to grvity. The pper film tends to move fster becse of the feeding of moistre from the rod bove it nd becse of the brekdown of srfce tension de to the rc. When moistre ddition is hevy, droplets tend to congregte nderneth the rod, which ccelertes this movement. In very hevy wetting conditions droplets rnning throgh the drybnd cn extingish the rc permnently or temporrily. A drybnd cn grdlly move down the rod by this mens ntil the lower wter film is replced by the immobile metllic fitting. The metllic fitting then becomes n immovble lower edge of the dry-bnd. In the mentime, the pper wter film cn still move downwrds. Following this, two possible sittions my occr: A) A single stble dry-bnd rc remins bt the rc length is physiclly compressed in length. The length of the rc is redced to extinction nless new blnce is chieved. B) The dry-bnd rc becomes nstble nd one or more drybnds re initited elsewhere on the rod s the originl drybnd is swmped nd the rc extingished. Sch movement of moistre cn reslt from previling winds on otdoor horizontl rod insltors (or ADSS cbles) [8]. However, for experimentl expediency nd control, the process hs been investigted sing sloping smples in the lbortory. Frther work will be reqired to consider the likelihood nd freqency of sch occrrences in service conditions. ) b) Upper wter film Wter film Stble Arcing Insltor core Wter film experiment to limit the lekge crrent. The dt cqisition system ws Ntionl Instrment Lbview system with voltge divider of,: to determine the voltge wveform cross the rod smple, nd mesrement resistor of kω to record crrent. The voltge nd crrent signls were stored in PC with smple rte of khz.. TEST PROCEDURE M obi l e A r c i ng Insltor core Lower wter film Figre. Dry-bnd rcing compression on insltor core between two wter films. I. The trnsformer voltge ws fixed t 7. kv throghot the tests. The smple ws set to the horizontl position in the slt-fog. It took bot n hor to obtin single stble rc on the mteril srfce. II. The smple ws lifted to. The length of rc ws redced reching eqilibrim fter mintes, t which time its new length ws recorded. Normlly the rc grdlly moved down the rod ntil it reched the lower electrode. III. The smple ws lifted to. mintes ws llowed for eqilibrim to be reched. Then the smple ws inclined to,,,, with mintes intervls between ech slope ngle, mesrements being tken ccordingly. IV. When the slope ngle reched, the rcing ws extingished. At this stge the lekge crrent wveform becme sinsoidl, in phse with the spply voltge. EXPERIMENTAL. TEST ARRANGEMENT Figre illstrtes the design of the inclined smple test. The mm long smple tested ws mm dimeter rod, consisting of mm rdil thickness commercil grde silicone rbber on glss-reinforced polymer core. A remotely controlled plley system lifts one end of smple to crete n incline. The silicone rbber srfce ws niformly roghened with brsive pper to redce its srfce hydrophobicity to encorge rpid formtion of continos wter films on the srfce nd so the initil formtion of stble dry-bnd rc. Electrodes, seprted by mm, were formed by wrpping copper strip tightly rond the rod. The wter film ws creted by the precipittion of srronding slt-fog throghot the test, with condctivity of 6, µs/cm, nd the precipittion rte in the chmber ws.±. l/m /hr. A wter resistor of MΩ ws sed throghot the Figre. The test rrngement.
. RESULTS nd crrent signls were obtined for slope increments between the horizontl nd, long with ssocited rc lengths. Figre gives exmples of test reslts. Detiled reslts re presented in Section. Arc intensity cn visibly be seen to increse s the rc is redced in length, becoming whiter in color. This is prticlrly evident for very short rcs s the rc is extingished, s the rod ngle is incresed bove. Arcing period - - - - - - - - - - Slope ngle: Eqilibrim rc length:. cm Arcing period - - Slope ngle: Eqilibrim rc length:.7 cm - - - - - - - - (ma) (ma) three periods re identified: the pre-rcing period, the rcing period nd the post-rcing period [6]. For simplifiction, the rc is represented s hving flt voltge dring its existence. The model is represented by the eqtions from these three periods s: Pre-rcing period: < t < t Arcing period: t < t < t i (t)= () (t)= Usinωt () i (t)= I sinω [t-( -t )] i ω () (t)=u () Post-rcing period: t < t < ms i (t)= () (t)= Usinω t (6) where: i (t) nd (t) re simlted crrent (ma) nd voltge (kv) crves. I nd ω i re the rms vle (ma) nd nglr freqency (rd/ms) of the crrent sinsoidl wve. U nd ω re the rms vle (kv) nd nglr freqency (rd/ms) of the voltge sinsoidl wve. U is the rc voltge (kv) dring the rcing period. t is the rc ignition time (ms). t is the rc extinction time (ms). Arcing period - - - - - - Slope ngle: Eqilibrim rc length:. cm Figre. Experimentl reslts of crrent nd voltge trces for slope ngles of, nd. - - - - (ma) U Pre-rcing Arcing Post-rcing t t c) Two sinsoidl model i(t) (crrent) (t) (voltge) Figre. Doble sinsoidl model to simlte I-V crves for dry-bnd rc. MODELLING. DOUBLE SINUSOIDAL MODEL A doble sinsoidl model cn be sed to nlyze the electricl chrcteristics of dry-bnd rcing dring the rccompression process. A limited model ws in introdced elsewhere [], nd this is now developed to llow vrition in pek crrent with rc length. In the model, two sinsoidl wves re introdced to simlte the experimentl crrent nd voltge chrcteristics, shown in Figre. In ech hlf-cycle. ARC CHARACTERISTICS WITH SLOPE ANGLE At ech slope ngle, the experiment dt from for hlfcycles ( points per hlf-cycle) were sed to nlyze rc length, rc period, brekdown voltge nd rc crrent pek. Figre shows the reltionship between rc length nd slope ngle. In this nd following figres, error brs indicte stndrd devitions of mesrement tken. For every hlf cycle, pictres were extrcted frme-by-frme from video imges nd Vistmetrix softwre sed to extrct the rc length
from the pictre. As rc length vries dring hlf cycle, the mximm rc length ws sed for nlysis. In the horizontl position ( slope), the rc grows withot externl forces moving the two wter films t the dry-bnd edge. Following the increse of slope ngle, the dry-bnd length ws redced. When the slope ngle reched rond, the dry-bnd re hs been entirely sbmerged by the pper wter film, extingishing the rcing ctivity. The rcing period is defined s the period from rc ignition time t to rc extinction time t. Figre 6 shows the experimentl reslts of rcing period ginst different rc lengths. As the rc is redced in length, the rcing period becomes longer. This is de to the chnge of rc ignition time with different rc length. Shorter dry-bnds reqire lower brekdown voltges, this in trn brings forwrd the time for rc ignition. However, the rc extinction time remins nchnged becse the rc extinction voltges re roghly the sme in ech slope ngle cse. Ths t is constnt prmeter in the model, independent of dry-bnd length nd rod inclintion. The brekdown voltge is defined here s the instntneos voltge t rc ignition. As the rc is compressed in length, the brekdown voltge redces linerly with dry-bnd length from to (Figre 7). The reson for this trend is tht dry-bnd rc compression reslts in shorter distnce of ir gp between the two wter lyers, reqiring lower threshold voltge to brek down the gp for rc ignition. The pek rcing crrent for ech dry-bnd length is shown in Figre 8. This reltionship shows tht the rc crrent pek rises linerly s rc length decreses. The extreme condition occrs t the slope ngle of, where the dry-bnd hs been swmped by the higher wter film rnning down the rod, leding to mximm lekge crrent, limited only by the externl circit s wter resistor. Arc Length (cm)... y = -.x +. R =.9 Slope ngle (deg) Figre. The reltionship between rc length nd slope ngle. Arcing period (ms) 9 8 7 6 y = -.986x +.7 R =.96...6.8.. Figre 6. The reltionship between rcing period nd rc length. Brekdown voltge (kv)...6.8.. Figre 7. The reltionship between brekdown voltge nd rc length. Arc crrent pek(ma).. y = -.76x +.868 R =.96... Figre 8. The reltionship between rc crrent pek nd rc length.. ARC PARAMETERISATION The reltionship between rc length, rcing period, rc ignition time nd rc extinction time is estblished empiriclly by nlyzing experimentl reslt s: T rc =t -t =-.99L +.6 (7) t =f(l,t )=t +.99L -.6 (8) where: T rc is the rcing period, t is the ignition time, t is the extinction time ( ll in ms) nd L is the rc length (cm). The reltionship between rc crrent pek nd rc length is represented s: I =.76L +.868 (9) I = (.76.868).7.99 L + = + L () where I is the rms vle (ma) of simlted crrent sinsoidl wve. A doble sinsoidl model of I-V crves for the dry-bnd rc compression in every hlf cycle is obtined by combining eqtions (-8) together with the sign fnction s follows: i (t)= I sinω [t-( -t )] i ω -sign[(t-t -.99L +.6)(t-t )] () +sign[(t-t -.99L +.6)(t-t )] (t)= Usinωt -sign[(t-t -.99L +.6)(t-t )] +U () where: i (t) nd (t) re simlted crrent (ma) nd voltge
(kv) crves, for < t < (ms). Tble smmrizes the prmeters for the doble sinsoidl model, estimted from experimentl dt. By observing the chnge in I-V crves with different rc lengths, it is determined tht the U, U, ω i, ω, nd t re constnt, while t, I nd L re vrible with dry-bnd length. t is represented by eqtion (8) nd I is represented by eqtion (). L vries representing different rc compression sittions.. SIMULATION RESULTS By introdcing the model prmeters from Tble to the model represented by eqtions (-), the doble sinsoidl model for dry-bnd rc compression consists of: i (t)=(-.9l +.) sin[.8(t-.)] {-sign[(t-.99l +.8)(t-8.88)]} () (t)=.sin(.t) {+sign[(t-.99l +.8)(t-8.88)]} +. {-sign[(t-.99l +.8)(t-8.88)]} () where: i (t) nd (t) re simlted crrent (ma) nd voltge (kv) crves for ms hlf cycle. i (t) nd (t) re mltiplied by - when < t < (ms). Figre 9 shows the simltion reslts of I-V crves eqivlent to reslts of Figre. ANALYSIS. COMPARISON BETWEEN SIMULATED AND MEASURED I-V TRACES Direct comprisons between simlted nd mesred I-V trces hve been mde to ssess the ccrcy of the doble sinsoidl model. Figre gives n exmple of grphicl comprison of mesred dt nd simltion for the.7 cm rc. The correltion coefficient, r, between mesrement nd simltion is clclted in ech cse nd presented in Tble. The correltion coefficients re consistently high, confirming tht the doble sinsoidl model is pproprite in these cses. Tble. Detils of model prmeters Doble Sinsoidl Model I-V crves [ma & kv] i (t) (t) Eqtion () () U [kv] 7. I [ma] I =f(l ) U [kv] 8.7 ω [rd/ms].. ω i [rd/ms].8 t [ms] t =f(l, t ) t =f(l, t ) t [ms] 8.88 8.88 L [cm].<l <..<L <. t [ms] <t< <t< Tble. Correltion coefficients of crrent nd voltge simltions. Arc length Correltion Coefficient (cm)..987.9.7.9.98..99.99 - - - - - - - - - - Simltion reslt Eqilibrim rc length:. cm - - Simltion reslt Eqilibrim rc length:.7 cm - - - - - - Simltion reslt Eqilibrim rc length:. cm Figre 9. Simltion reslts of crrent nd voltge trces. (ma) - - - - - - - - - - Simlted voltge Mesred voltge Eqilibrim rc length:.7 cm Simlted crrent Mesred crrent Eqilibrim rc length:.7 cm Figre. An exmple of comprison between simlted nd mesred I-V trces for n rc length of.7 cm - - - - - - - - - - - - (ma) (ma) (ma)
. ARC RESISTANCE Instntneos rc resistnce cn be obtined by the rtio of rc voltge to rc crrent. Experimentlly this is obtined from the mesred voltge nd crrent trces. In the doble sinsoidl model, the instntneos rc resistnce (MΩ), r (t), for times t <t<t is simlted by sing the (t) nd i (t) of eqtions () nd (), so tht dring the rcing period: (t) r (t)= = i (t) U I sinω [t-( -t )] i ω () By sbstitting the relevnt model prmeters from Tble we obtin, r (t)= i (t) (t) 8.7 = (-.8L +.8)sin[.8(t-.)] (6) Figre shows exmples of both experiment nd simlted rc resistnces in different rc compression (dry-bnd length) sittions. One chrcteristic of instntneos rc resistnce is its minimm vle. This minimm resistnce is locted on the bottom of the U shpe in Figre where the rcing crrent is gretest. Figre smmrizes the minimm rc resistnce ginst slope ngle for both experiment nd simltion. Trend lines indicte tht the more inclined the smple, the lower the rc resistnce becomes. This reslts from the redction of drybnd length: when the rc hs been more compressed, the rc crrent pek increses (Figre 8) bt the rc voltge does not chnge significntly with crrent. Resistnce (MΩ) Figre. The instntneos rc resistnce from both experimentl mesrement nd simltion work, for one hlf cycle ( ms). Minimm rc resistnce (MΩ) Slope L.9 cm Simlted R Mesred R 6 Slope L.7 cm Simlted R Mesred R Experimentl trend Liner (Experimentl trend) Slope L. cm Simlted R Mesred R Slope ngle (deg) Simltion trend Liner (Simltion trend) Slope Arc extinction Mesred R Figre. The reltionship between minimm rc resistnce nd slope ngle in both experimentl cse nd simltion cse. Resistnce per nit length (MΩ/cm) 6 Experimentl trend Simltion trend Slope ngle (deg) Figre. The reltionship between minimm rc resistivity nd slope ngle. The rc resistnce per nit rc length lso vries dring the rcing period s U shped crve. The minimm vle per hlf cycle s fnction of slope ngle is smmrized in Figre. From both experiment nd simltion, it is fond tht nlike the trend of resistnce, the minimm rc resistnce per nit length does not drop with the rc compression. Clerly significnt improvement in the model of instntneos resistnce cold be obtined by introdcing non-constnt rcing voltge (insted of U ). However, since the object of this stdy is the therml chrcteristics of the rc, which re essentilly verged over repeted cycles, this ws considered not to be necessry.. ARC POWER nd ENERGY Arc power is significnt chrcteristic which my led to dmge of insltor mteril. Bsed on the mesred I-V trces, the rc powers of ech hlf-cycle corresponding to different slope ngles re clclted nd shown in Figre. The reslt shows tht the pek power for different rc lengths stys rond W nd does not chnge too mch. However, the rcing period increses with rod inclintion. For the experimentl reslts, ech hlf cycle of dry-bnd rcing hs 8 mesrement points. Bsed on these discrete dt of rc power, the rc energy is clclted s: 8 n n+ n (7) n= E = [P(t ) + P(t )] t / where E is the mesred rc energy for hlf cycle, P(t n ) is the rc power (W) t the n th smple point, nd t n is the intervl smple time (. ms). Arc power (W) 8 6 - - -6-8 -...9.6.77 Arcing Period Slope Angle 6.7 6. 8. Figre. Arc power of different slope ngles with vrible rcing period.
Bsed on the simlted I-V wves, the ccmlted rc energy for hlf cycle of dry-bnd rc is clclted s: E = (t)i (t)dt (8) where: E is the simlted rc energy (Jole) per hlf cycle, i (t) is the simlted crrent wve, nd (t) is the simlted voltge wve. In both the pre-rcing ( < t <t ) nd post-rcing (t <t < ) periods i (t) is zero, which mens there is no power dissiption dring these two periods. As reslt, the eqtion for rc energy clclted cn be simplified s: t E = { U I sinω [t-( -t )]}dt (9) i ω t Sbstitting the model prmeters from Tble into eqtion 9, the simlted rc energy for hlf cycle of drybnd rcing is given by: E =(-.8L +.6) [+cos(-.9+.l )] () The rc energy from the experimentl reslts sing eqtion (7) nd the simltion reslt of eqtion () re smrized nd compred in Figre. In the sittion of dry-bnd being compressed in length, the redction of rc length reslts in rise in rc energy per hlf cycle. The energy density, D, for ech hlf cycle of dry-bnd rc cn be clclted be dividing rcing energy by the dry-bnd re over which rcing occrs: E D= (Jole/m ) () RL where R is the rdis of smple rod, L is the rc length, nd RL is the dry-bnd re. Figre 6 demonstrtes the chnge of rc energy density with rc length from experimentl nd simltion reslts. The energy density increses by fctor of 6 with the redction of rc length from mm to mm. This is de to the simltneos rise of rc energy nd the redction of dry-bnd re dring the rc compression. Arc energy (Jole)...9.8.7.6.. Simltion reslt...7.. Figre. Experimentl nd simltion rc energy ginst rc length s reslt of rc compression. Energy density (Jole/m) 8 6 Simltion reslt...7.. Figre 6. The reltionship between rc length nd energy density chnge. CONCLUSIONS An experimentl stdy hs been condcted to investigte the dry-bnd rcing on rod geometries, nd hs shown tht stble rcs cn be compressed in length by chnging the ngle of the rod to the horizontl. Sch rcs remin stble nd hve been redced in length by over %. The mesred voltge nd crrent wveforms hve been modeled sing simple doble sinsoidl model with constnt rc voltge. A notble fetre of this model is tht the time of extinction of the rcs dring the power freqency cycle ws independent of rc length. As the dry-bnd nd its ssocited rc re redced in length the drtion of the rc dring ech hlf cycle increses, s the brekdown voltge of the dry-bnd is redced. The minimm resistnce of the rc is lso redced with the rc length leding to higher pek crrent s the rc length is diminished. As reslt rc compression leds to higher rc energy, nd in prticlr higher rc energy per nit re of dry-bnd. In the cse stdied here the rc energy per nit dry-bnd re ws incresed by fctor of s reslt of hlving its length. It is sggested tht sch dry-bnd rc compression processes my led to more ggressive dmge on the composite insltor srfce, nd cold possibly ccelerte the long-term ging effect into short-term hzrd. As reslt, processes controlling insltion lifetime my not be continl nd grdl, bt re determined by events sch s the occrrence of dry-bnd rc compression. Sch event my be de to dventitios environmentl circmstnces, for exmple the combintion of high polltion levels nd wind. Frther stdy is reqired of the short-term nstble trnsient rc extinction process, since extrpoltion of the rc length power crve sggests rpidly incresed energy density s the length redces frther beyond the stble conditions exmined here. Moreover, geometriclly the rc is closer to the rod srfce s it becomes shorter, thereby frther enhncing the trnsfer of power to the mteril perhps leding to even greter thret to the srfce integrity. ACKNOWLEDGMENTS The thors wold like to thnk Ntionl Grid for their spport in this work, nd their permission to pblish.
REFERENCES [] R. S. Gorr, E. A. Cherney nd J. T. Brnhm, Otdoor Insltors, ISBN -96776--7, Phoenix, Arizon, USA, 999. [] Y. Xiong, S. M. Rowlnd, J. Robertson nd R. J. Dy, "Srfce Anlysis of Asymmetriclly Aged kv Silicone Rbber Composite Insltors", IEEE Trns. Dielectr. Electr. Insl., Vol., pp. 76-77, 8. [] S. H. Kim, E. A. Chemey nd R. Hckm, "Effect of Dry Bnd Arcing on the Srfce of RTV Silicone Rbber Cotings", IEEE Intern. Sympos. Electr. Insl., Bltimore, MD, USA, pp. 7-, 99. [] Y. Zh, M. Otsbo, N. Anmi, C. Hond, O. Tkenochi, Y. Hshimoto, nd A. Ohono, "Chnge of Polymeric Mteril Exposed to Dry Bnd Arc Dischrge [Polymer Insltor Applictions]", IEEE Conf. Electr. Insl. Dielectr. Phenomen (CEIDP), pp. 6-68,. [] M. Kmos, L. Kmos nd D. Armentrot, "Filre Anlyses of Noncermic Insltors Prt : Brittle Frctre Chrcteristics", IEEE Electr. Insl. Mg., Vol., No., pp. -7,. [6] S. Kmgi nd N. Yoshimr, "Trcking nd Erosion of HTV Silicone Rbber nd Sppression Mechnism of ATH", IEEE Trns. Dielectr. Electr. Insl., Vol. 8, pp. -,. [7] S. M. Rowlnd nd I. V. Nichols, "The Effects of Dry-Bnd Arc on Ageing of Self-Spporting Dielectric Cbles in High Fields", IEE Proc.- Sci. Mes. Technol, Vol., pp. -, 996. [8] S. M. Rowlnd nd F. Esthope, "Electricl Ageing nd Testing of Dielectric, Self-Spporting Cbles for Overhed Power Lines", Proc. IEE Prt A, Vol., pp. -6, 99. [9] G. G. Krdy, S. Bozhng, H. Qi, D. Srinivsn nd M. W. Tominen, "Experimentl Investigtion of the Aging Process on ADSS Opticl Fiber Cbles", IEEE Power Engineering Soc. Smmer Generl Meeting, pp. -8,. [] S. Bozhng, G. G. Krdy, H. Qi nd M. W. Tominen, "Experimentl Stdies of the Chrcteristics of Dry Bnd Arcing on ADSS Fiber Optic Cbles", IEEE Trns. Power Delivery, Vol. 9, pp. 96-9,. [] S. Bozhng, H. Qi, G. G. Krdy nd M. W. Tominen, "Stdies on the Length of Dry Bnd Arcs on ADSS Fiber Optic Cbles", IEEE Conf. Electr. Insl. Dielectr. Phenomen (CEIDP), pp. -6,. [] H. Qi, S. Bozhng, G. G. Krdy nd M. Tominen, "Stdy on the Effect of Test Circit Prmeters on Dry Bnd Arcing Filre of ADSS Fiber Optic Cble", IEEE Power Engineering Soc. Smmer Generl Meeting, pp. 7-,. [] Q. Hng, G. G. Krdy, B. Shi nd M. Torninen, "Clcltion of the Electric Field Distribtion on ADSS Fiber Optic Cble", IEEE Conf. Electr. Insl. Dielectr. Phenomen (CEIDP), pp. 79-8, [] L. H. Meyer, S. H. Jyrm nd E. A. Cherney, "Correltion of Dmge, Dry Bnd Arcing Energy, nd Tempertre in Inclined Plne Testing of Silicone Rbber for Otdoor Insltion", IEEE Trns. Dielectr. Electr. Insl., Vol., pp. -,. [] X. Zhng, S. M. Rowlnd nd V. Terzij, "Modelling of Dry-Bnd Arc Compression", Accepted for pbliction in 6th Interntionl Symposim on High Engineering, 9. [6] H. Qi, G. G. Krdy, B. Shi nd M. Tominen, "Nmericl Simltion of Dry-Bnd Arcing on the Srfce of ADSS Fiber Optic Cble", IEEE Trns. Dielectr. Electr. Insl., Vol., pp. 96-,. Xin Zhng ws born in Jinn, Chin. He completed the B.Eng. degree in tomtion t Shndong University in 6, nd received the M.Sc. degree with distinction in electricl power engineering from the University of Mnchester in 7. He is crrently stdying for his Ph.D. degree in electricl power engineering t the Ntionl Grid High Reserch Center in the sme University. Simon M. Rowlnd (SM 7) ws born in London, Englnd. He completed the B.Sc. degree in physics t UEA nd the Ph.D. degree t London University. He ws wrded the IEE Dddell Premim in 99 nd becme FIEE in. He hs worked for mny yers on dielectrics nd their pplictions. He hs lso been Opertions nd Technicl Director mltintionl mnfctring compnies. He joined The School of Electricl nd Electronic Engineering in The University of Mnchester s Senior Lectrer in nd ws ppointed Professor of Electricl Mterils in 9. He is Vice-President Admin. of the IEEE Dielectric nd Electricl Insltion Society. Vldimir Terzij (M 9-SM ) is the EPSRC Chir Professor in Power System Engineering in the School of Electricl nd Electronic Engineering, The University of Mnchester, where he hs been since 6. From 997 to 999, he ws n Assistnt Professor t the University of Belgrde. In 999, he ws wrded prestigios Hmboldt Reserch Fellowship. From to 6, he ws with ABB AG, Germny, s n expert for switchgers nd distribtion tomtion. His min reserch interests re ppliction of intelligent methods to power system monitoring, control, nd protection, switchger, nd DSP pplictions in power systems.