Korona

2518 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 28, NO. 4, OCTOBER 2013

Lightning Impulse Corona Characteristic of 1000-kVUHV Transmission Lines and Its Influences onLightning Overvoltage Analysis ResultsPengcheng Yang, Shuiming Chen, Senior Member, IEEE, and Jinliang He, Fellow, IEEE

AbstractThe lightning overvoltage is one of key issues in de-signing the insulation of transmission line and power apparatus.Usually the influence of the impulse corona is ignored in numer-ical analysis of electromagnetic transients in power systems dueto the lack of experimental data. But the relationship between theovervoltage and the insulation size of 1000-kV ac ultra-high voltage(UHV) power systems enters the nonlinear region, and a small de-crease in overvoltage will lead to a high reduction in insulationsize and bring huge economical benefits. So fully understandingthe corona characteristics of the 1000-kV UHV transmission lineis helpful to understand the actual wave process and realize a suit-able insulation design of the 1000-kVUHV system. Experiments onthe lightning impulse corona characteristic of bundle conductorsfor 1000-kV transmission lines are carried out in the China UHVac test base and the impulse corona Q-V curves of different bundleconductors are obtained. The results show that the impulse coronaleads the increment of the capacitance of the conductor. Analysison the Q-V curves shows that the occurrence of the positive im-pulse corona has a time delay after the lightning impulse voltageexceeds the corona inception voltage and the delay time is a tenthof a microsecond. At the end of this paper, a simulation model forcorona based on the empirical formulas of Q-V curves fitted fromthe test results is proposed, and is applied to the overvoltage anal-ysis of 1000- kV transmission line and substation. The analysis re-sults show the lightning impulse corona has a large influence on thelightning overvoltage of the transmission line and substation.

Index TermsCorona cage, electromagnetic transient calcula-tions, impulse corona, lightning impulse, Q-V curve, ultra-highvoltage (UHV) bundle conductors.

I. INTRODUCTION

L IGHTNING overvoltage plays an important role in de-signing the insulation coordination of transmission linesand substations. In order to simulate the lightning overvoltagesprecisely, many studies have shown how to consider the light-ning impulse corona [1][6]. The corona discharge around theconductors decreases the conductor impedance and increases

Manuscript received April 08, 2013; revised July 05, 2013 and August 03,2013; accepted August 07, 2013. Date of publication August 30, 2013; date ofcurrent version September 19, 2013. This work was supported in part by theNational Natural Science Foundation of China under Grant 5073001, in partby the National Basic Research Program of China (973 Project) under Grant2009CB724504), and in part by the China State Power Grid. Paper no. TPWRD-00408-2013.The authors are with the Department of Electrical Engineering, State Key

Lab of Power Systems, Tsinghua University, Beijing 100084, China (e-mail:[email protected]).Color versions of one or more of the figures in this paper are available online

at http://ieeexplore.ieee.org.Digital Object Identifier 10.1109/TPWRD.2013.2278120

Fig. 1. Typical impulse corona curve.

the coupling factors between the lines. Thus, the wave distor-tion and attenuation caused by impulse corona in propagationshould be considered in the precise simulations.When the voltage on the conductor exceeds the corona in-

ception voltage, the charges are generated not only on the con-ductor surface but also in the adjacent space [7]. Normally, thecharge versus voltage ( - ) curve of the conductor is used torepresent the characteristic of impulse corona [8][10]. A typ-ical curve as shown in Fig. 1, consists of three parts [8].For a voltage of less than is relative to voltage, and theslope of the curve is the natural capacitance of the con-ductors . For voltages greater than the corona inceptionvoltages and less than increases nonlinearly withslope , which is larger than . For the voltagesdecreasing section, the charge on the conductor decreases withthe voltage but the space charge scarcely decreases. As a result,the slope of the curve is close to in this section.Many experimental studies have been carried out on the light-

ning impulse corona characteristics of conductors in the coronacage [12], [13] or the conductors above a ground plane [7],[14][18]. However, only several experiments have been per-formed on the impulse corona characteristics of 1000-kV ultra-high voltage (UHV) conductors under positive and negativelightning impulses [18], [19]. These experimental results are notenough to establish precise impulse corona model in the simu-lations. As we know, the relationship between the overvoltageand the insulation size enters the nonlinear region in UHV sys-tems [20], a small decrease in overvoltage will lead to a high de-crease in insulation size and bring huge economic benefits. So tofully understand the corona characteristics of the 1000-kVUHVtransmission line, it is helpful to understand the actual waveprocess and realize a suitable insulation design of the 1000-kVUHV ac power system.

0885-8977 2013 IEEE

YANG et al.: LIGHTNING IMPULSE CORONA CHARACTERISTIC OF 1000-kV UHV TRANSMISSION LINES 2519

Fig. 2. Experimental setup.

This paper presented the experiments on lightning coronacharacteristics of single and bundle conductors under positiveand negative lightning impulses. The tests were conducted inthe outdoor corona cage in the China UHV AC Test Base inWuhan, China. Fourteen kinds of bundle conductors, which arecommonly used in 110- to 1000-kV transmission lines in China,were tested in the experiments, and the lightning impulse corona- characteristic curves of different bundle conductors were

obtained. Based on the test results, empirical formulas of -curves were fitted and an impulse corona model was establishedfor the electromagnetic transient analysis. The effect of impulsecorona on the 1000-kVUHV system lightning overvoltages wasdiscussed at the end of this paper.

II. EXPERIMENTAL SETUP

A. Test Equipment

The test equipment is comprised of a corona cage, a4800-kV/720-kJ impulse voltage generator, an RC dividerconnected on the tested conductors, and a charge measurementcapacitor connected between the internal and external cages.The experimental setup is shown in Fig. 2.The metallic-wire mesh cage consists of a 25-m-long central

measuring section, and two grounded shielding sections that areeach 5 m long. The conductors to be tested are suspended in theinternal cage with a cross section of 8 8 m. Sprinklers areprovided at regular intervals along the cage to produce artificialrain at a rate of 50 mm/h. The corona cage is equipped withthree sets of wire connection hardware to suspend 1 to 12 bundleconductors.

B. Measurement of the Q-V Curve

The measurement of the impulse characteristic requiresa simultaneous recording of charge ( ) and voltage ( ) signals,so the precision of the two transient recording devices plays akey role on the accuracy of the curve. A four-channel,100-MHz, and 16-b data-acquisition system is used to recordthe signals.The impulse voltage is measured by the voltage divider with

6 capacitors in series, the divider ratio is 4747, and the measure-ment uncertainty is less than 2%. The space charge is measuredby the capacitor connected between the internal and external

Fig. 3. Measurement capacitor connected between the internal and externalcages.

Fig. 4. Measurement system of the impulse corona curve.

cages, as shown in Fig. 3. Since the duration of impulse is veryshort, space charges are generated on the conductor surface andthe adjacent space. The space capacitance between the conduc-tors and the internal cage is in series with the measurement ca-pacitor. The charge on the measurement capacitor approximatesto the one generated by the impulse corona. In the experiments,the measurement capacitor is 1.006 F and the space capaci-tance between the conductors and the internal cage is 0.014 F.The charges could be derived by the voltages over these capac-itors. The charge (Q) and voltage (V) signals are transmitted tothe data acquisition by the coaxial cable with the same lengthsimultaneously. The measurement system is shown in Fig. 4.

III. TESTS RESULTS

As shown in Table I, 14 kinds of conductors were tested. Eachkind of conductor was tested under both positive and negative2.6/50- s lightning impulse voltage.Presently, 10- and 12-bundle conductors are used in Chinas

1000-kV ac UHV single-circuit and double-circuits transmis-sion lines, respectively. The corresponding curves areshown in Figs. 5 and 6. curves of other bundle conductorsare shown in the Appendix.

2520 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 28, NO. 4, OCTOBER 2013

TABLE ICONDUCTORS TESTED IN THE IMPULSE CORONA EXPERIMENT

Fig. 5. Lightning impulse corona curves of 8 LGJ500/35 s400:(a) Negative. (b) Positive.

Fig. 6. Lightning impulse corona curves of 8 LGJ630/45 s400:(a) Negative. (b) Positve.

IV. DISCUSSIONS

A. Corona Inception VoltageThe corona inception voltage is the principal characteristic of

the curve. When the impulse voltage exceeds the coronainception voltage, charges are generated in the space around theconductors. It can be seen that the curve deviates from the nat-ural capacitance and increases rapidly. Thus, the knee point ofthe curve is usually considered as the conductors coronainception voltage.As shown in Figs. 5 and 6, the curves become steeper at the

knee points. The steepness of the positive curves is much greaterthan the negative ones. The phenomenon can be explained bythe gas discharge. In the negative corona cases, a large numberof free electrons are generated and attached to the air moleculesaround the conductors. The negative ions increase the conductorequivalent radius and restrain the further development of the

TABLE IIPOSITIVE LIGHTNING IMPULSE CORONA INCEPTION DELAY

TIME OF UHV CONDUCTORS

corona. In the positive corona cases, positive streamers are gen-erated on the surface of the conductors, and the electric fieldin front of the streamers is strengthened. As a result, a coronaunder positive lightning impulse voltage is much easier to gen-erate. The positive corona is much stronger than the negativeone. It can be seen from the curves in Figs. 5 and 6 that, underthe same impulse voltage, the positive corona generates morecharges than the negative one does.The knee point is usually considered as the corona inception

voltage. However, in the positive corona curves, the kneepoints of the same conductors are not kept the same, when theamplitudes of the impulse voltage applied are different andthe other parameters remain unchanged. This phenomenonhas never been observed in the negative lightning impulse andswitching impulse corona.The impulse corona inception delay time may explain the

phenomenon. The positive impulse corona will not be generatedat the time the voltage exceeds the corona inception voltage.The reason is that the motion velocity of the positive ions is ob-viously slower than that of the negative electrons, and the pos-itive ions on the conductor surface need a very short period tomove into space to form positive charges. But for negative im-pulse corona, the negative electrons are much faster to moveinto space to form negative charges. It is possible that there isa delay time in the negative impulse as well, but the negativeimpulse corona inception delay time is much shorter than thelightning wavefront time and has little influence on the curve.The knee points of the UHV conductors are obtained from the

impulse voltage waveform recorded. Table II shows the positivelightning impulse corona inception delay time of UHV conduc-tors. As shown in Fig. 1, there is a delay time between the coronainception voltage and the knee point. For the 8 LGJ500/35bundle conductors, the delay time is from 0.45 to 0.73 s. Forthe 8 LGJ630/45 conductors, the delay time is about 0.5 s.The delay time of the other conductors ranges from 0.4 to 0.9s without obvious regularity. Under the lightning impulse, thevoltages rise rapidly in the delay time period and the voltage ofthe knee point is much higher than the corona inception voltage.The steeper the impulse is, the higher the knee point voltage is.The knee point voltages of the same conductors under differentamplitude impulses are scattered over a wide range.The wavefront time of switching impulse is several hundreds

of microseconds [21], [22], three orders of magnitude largerthan the delay time. The switching impulse voltage barelychanges during the delay time. That explains the positive

YANG et al.: LIGHTNING IMPULSE CORONA CHARACTERISTIC OF 1000-kV UHV TRANSMISSION LINES 2521

Fig. 7. Lightning impulse corona curve of 8 LGJ630/45 s400 underrainy conditions: (a) Negative. (b) Positive.

Fig. 8. Positive lightning impulse corona curve of 8 LGJ630/45 s400bundle conductors under fair and rainy conditions.

switching impulse of the same conductor having the same kneepoint. Thus, in the switching impulse corona test, the knee pointcould be considered as the corona inception voltage. However,under lightning impulse, the voltages change greatly during thedelay time. As a result, it is difficult to identify the lightningcorona inception voltage by knee point. This delay time in theswitching impulse corona is a tenth of a microsecond, whichcan be ignored.

B. Effect of Weather on Lightning Corona

Lightning stroke on transmission lines is usually accompa-nied with raining. The lightning corona characteristic underrainy conditions has drawn great attention. Tests on the UHVbundle conductors under the condition of artificial rain werecarried out. Fig. 7 shows the curve of 8 LGJ630/45bundle conductors under a rainy condition.Fig. 8 compares the positive lightning impulse corona

curves under fair and rainy conditions. The keen point underthe rainy condition is about 900 kV and the keen point undersunny conditions is 970 kV, so it is obvious that the corona onsetvoltage under rainy conditions is much lower than that under thesunny condition, and the corona generates more charges underthe rainy condition. Therefore, the impulse corona loss is greaterunder the rainy condition, as shown in Fig. 9. While there islittle difference between the negative lightning impulse and thenegative impulse corona, curves under rainy conditionsalmost overlap those under fair conditions.

Fig. 9. Impulse corona loss under sunny and rainy conditions.

V. LIGHTNING CORONA MODEL FOR ELECTROMAGNETICTRANSIENT CALCULATIONS AND ITS APPLICATIONS

A. Lightning Corona Model

Based on the results obtained from the experiments, the em-pirical formulas of curves can be summarized. Due tothe time delay in the positive corona inception, it is difficultto propose a uniform equation for positive impulse presently,and more work should be done. The negative lightning impulsecorona curve of the bundle conductors can be calculatedby

(1)

(2)

where is the corona inception voltage, and is the coefficientof conductors. The natural capacitance of conductorscould be calculated by finite-element analysis software or by thefollowing approximate formula:

linescages (3)

where is the subconductor radius, is the radius of thebundle conductors. is the height of transmission lines, and

is the radius of the corona cage.The corona inception voltage can be calculated by

where is the corona inception gradient according to Peeksformula. is the relative air density and is theconductor surface roughness factor.

2522 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 28, NO. 4, OCTOBER 2013

TABLE IIICOEFFICIENTS IN THE CURVE OF DIFFERENT BUNDLE CONDUCTORS

UNDER NEGATIVE LIGHTNING IMPULSES

Fig. 10. Lightning impulse corona model.

The coefficients of different bundle conductors under nega-tive lightning impulse are listed in Table III.Based on the empirical equations of curves, an appli-

cable corona model is shown in Fig. 10. Transmission linesare divided into several sections, and the nonlinear capacitancecircuits are added in the electromagnetic transient programs[2], [9], [23][25], which is different from the linear elementcorona model proposed by Motoyama [26], based on Leescorona model. If the number of discretizations of the transmis-sion line is not enough, the corona models inserted may causewave deformation, resulting in reduced overvoltage. This hadbeen discussed in [25] and the section of the line was decreasedfrom 200 to 100 to 50 m. Substantial differences were apparentin the results from 200 to 100 m, but the differences were smallfrom 100 to 50 m. Thus, in this paper, the transmission linewas divided into sections of 50 m to obtain precise simulationresults. When the applied impulse voltage is lower than thecorona inception voltage , the diode is turned off and the linecharge increases linearly with voltage . The slope of the curve is the conductor natural capacitance .

If the corona inception voltage is exceeded, the diode is turnedon and additional capacitance is connected with thecircuit in parallel to simulate the effect of corona. When thecrest voltage is exceeded, the diode is turned off andonly remains in the circuit. The charge decreasesfollowing the slope equal to . can be calculatedby the empirical equations and is changed in every time stepwith the line voltages in the transient simulations.The energy dissipation during the corona process can be rep-

resented by a conductance that comes into effect only at volt-ages above corona onset, and this could make the model muchmore accurate. However, the duration of lightning impulse is

Fig. 11. Structure of the 1000-kV UHV double-circuit transmission lines inChina.

Fig. 12. Overvoltage waveforms on the insulator string when a negative light-ning strikes the upper phase line of the 1000-kV transmission line.

very short and the energy dissipation is very small, so energydissipation is not considered in this paper.

B. Application in Lightning Overvoltage Analysis of the UHVTransmission LineTaking the 1000-kV UHV double-circuit transmission line in

China as an example, the effect of lightning impulse corona iscalculated. Fig. 11 shows the tower structure of the UHV trans-mission line. The transmission-line model was established inPSCAD and the line is frequency dependence. When a negativelightning current of 30 kA strikes the upper phase line, the light-ning overvoltage on the insulator is compared with that withoutconsidering the lightning impulse corona in Fig. 12.As shown in Fig. 12, the amplitude of the impulse voltage on

the insulator with the consideration of corona is 9% less thanthat without the corona. The lightning impulse withstand levelfor shielding failure rises from 33 to 37 kA, which means a12.1% increase.

C. Application in Lightning Overvoltage Analysis of the UHVSubstationThe 1000-kVUHV substation connected with double-circuits

transmission lines in China is considered in analysis. The UHVtransmission line connected with the UHV substation is morethan 150 km, and the distance between the first tower and the

YANG et al.: LIGHTNING IMPULSE CORONA CHARACTERISTIC OF 1000-kV UHV TRANSMISSION LINES 2523

Fig. 13. V-I characteristic of surge arresters.

Fig. 14. Diagram of the 1000-kV substation.

substation might be 200 m. The lightning striking point is only1 km away from the substation; thus, the wave deformationcaused by frequency dependence of the transmission line can beignored. The corona model in Fig. 10 was used in the lightingovervoltage calculations and the characteristic of the surgearresters is shown in Fig. 13.Fig. 14 is the diagram of the 1000-kV substation. Since the

insulator flashover voltage of the UHV system is very high, thelightning shielding failure is the decisive factor on the designof equipment insulations in the substations. If 50-kA lightningstrikes the 1000-kV UHV transmission line 1 km away fromthe substation, the overvoltage on the transmission line is largerthan 6000 kV and causes strong impulse corona.Fig. 15 shows the calculation results of the overvoltages at

the first tower (which is 200 m away from the substation) ofthe 1000-kV UHV substation. If the impulse corona is ignoredin the calculation, the amplitude of the lightning overvoltageis 7500 kV and the wavefront time is 1.35 s, whereas if theimpulse corona is considered, the amplitude decreases to 5500kV and the wavefront time increases to 1.69 s. Furthermore,the overvoltages on the equipment of the substation are shownin Table IV.As shown in Table IV, the lightning overvoltages on the

equipment are quite different in the two calculation conditions.With the consideration of impulse corona, the overvoltages onthe series reactor decrease to 1731 kV from 2035 kV with thedecrease rate of 14.9%, lower than the BLIW 1957 kV. Thus,it is important to consider the impulse corona effect in thelightning overvoltage calculations. In this case, if the impulse

Fig. 15. Influence of impulse corona on the overvoltages at the first tower(which is 200 m away from the substation) of the 1000-kV UHV substationwhen 50-kA lightning strikes the transmission line 1 km away from thesubstation.

Fig. 16. Lightning impulse corona curves of LGJ630/45: (a) Negative.(b) Positive.

Fig. 17. Lightning impulse corona curves of 4 LGJ630/45 s450:(a) Negative. (b) Positive.

Fig. 18. Lightning impulse corona curves of 6 LGJ630/45 s450:(a) Negative. (b) Positive.

corona is not considered, a set of surge arresters has to beinstalled on the series reactor; but when the impulse coronais considered in the analysis, there is no need to install surgearresters to protect the series reactor, and unnecessary cost isavoided.

2524 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 28, NO. 4, OCTOBER 2013

Fig. 19. Lightning impulse corona curves of 10 LGJ630/45 s350:(a) Negative. (b) Positive.

TABLE IVLIGHTNING OVERVOLTAGES ON THE EQUIPMENT

OF THE 1000-kV UHV SUBSTATIONS

VI. CONCLUSION

The knee point of the curve is usually considered tobe the corona inception voltages. However, under the positivelightning impulse, the knee points of the same conductors aredifferent. The voltage at the knee point is much larger than thecorona inception voltage. This phenomenon had appeared inlightning impulse corona test results in the literature [19]; how-ever, researchers had not made a systematic study on it. Coronainception delay time is proposed in this paper to explain thisphenomenon.The positive impulse coronas are not generated at the time the

voltage exceeds the corona inception voltage. After a very shortperiod, the charges are generated in the space. Analyses on thetest results show that the delay time of positive corona rangesfrom 0.4 to 0.9 s without obvious regularity. This delay timeis a tenth of a microsecond that can be ignored in the switchingimpulse corona. However, it has a significant influence on thelightning impulse corona characteristic.The weather conditions have some effect on the lightning im-

pulse corona characteristic. For positive impulses, the coronaonset voltage is much lower and the corona loss is greater underrainy conditions. While for negative impulses, the corona curves under rainy conditions almost overlap those under fairconditions.Based on the test results, the empirical formulas of

curves are developed. A model to simulate the corona effect ofconductors in the transient calculation is established. The light-ning overvoltages on the UHV tower and the UHV substationsare calculated. In our case, with the effect of impulse corona,the overvoltage on the 1000-kV insulator decreases by 9%, andthe tower lightning impulse withstand level for shielding failurerises from 33 to 37 kA. The impulse corona could lead the greatattenuation and distortion to the overvoltage waveshape at the

incoming point of the 1000-kV UHV substation, and the ampli-tude decreases by 27% and the wavefront time does by 25%. Inthe meantime, the overvoltages on the equipment of the UHVsubstation are much lower when impulse corona is considered,for example, the overvoltages on the series reactor decreasefrom 2035 to 1731 kV, lower than the BLIW 1957 kV; as a re-sult, there is no need to install extra surge arresters to protect theseries reactor, and unnecessary cost is avoided.

APPENDIX A

The curves of other conductors are shown inFigs. 1619

ACKNOWLEDGMENT

The authors would like to thank Prof. H. Weigang for thesuggestions on the tests plan. The authors also appreciate thefield tests and equipment supported by W. Xiong, X. Tao, andother staff at the UHV AC Test Base.

REFERENCES[1] V. Cooray and N. Theethayi, Pulse propagation along transmission

lines in the presence of corona and their implication to lightning returnstrokes, IEEE Trans. Antennas Propag., vol. 56, no. 7, pp. 19481959,Jul. 2008.

[2] H. M. Kudyan and C. H. Shih, A nonlinear circuit model for trans-mission lines in corona, IEEE Trans. Power App. Syst., vol. PAS-100,no. 3, pp. 14201430, Mar. 1981.

[3] M.Mihailescu-Suliciu and I. Suliciu, A rate type constitutive equationfor the description of the corona effect, IEEE Trans. Power App. Syst.,vol. PAS-100, no. 8, pp. 36813685, Aug. 1981.

[4] T. H. Thang, Y. Baba, N. Nagaoka, A. Ametani, J. Takami, S. Okabe,and V. A. Rakov, A simplified model of corona discharge on anoverhead wire for FDTD computations, IEEE Trans. Electromagn.Compat., vol. 54, no. 3, pp. 585593, Jun. 2012.

[5] T. H. Thang, Y. Baba, N. Nagaoka, A. Ametani, J. Takami, S. Okabe,and V. A. Rakov, FDTD simulation of lightning surges on overheadwires in the presence of corona discharge, IEEE Trans. Electromagn.Compat., vol. 54, no. 6, pp. 12341243, Dec. 2012.

[6] C. de Jesus and M. T. Correia de Barros, Modelling of corona dy-namics for surge propagation studies, IEEE Trans. Power Del., vol.9, no. 3, pp. 15641569, Jul. 1994.

[7] T. Noda, T Ono, H. Matsubara, H. Motoyama, S. Sekioka, and A.Ametani, Charge-voltage curves of surge corona on transmission linestwo measurement methods, IEEE Trans. Power Del., vol. 18, no. 1,pp. 307314, Jan. 2003.

[8] A. Inoue, Propogation analysis of overvoltage surges with coronabased upon charge versus voltage curve, IEEE Trans. Power App.Syst., vol. PAS-104, no. 3, pp. 655662, Mar. 1985.

[9] P. S. Maruvada, D. H. Nguyen, and H. Hamadani-Zadeh, Studies onmodeling corona attenuation of dynamic overvoltages, IEEE Trans.Power Del., vol. 4, no. 2, pp. 14411449, Apr. 1989.

[10] X. R. Li, O. P. Malik, and Z. Zhao, A practical mathematical model ofcorona for calculation of transients on transmission lines, IEEE Trans.Power Del., vol. 4, no. 2, pp. 11451152, Apr. 1989.

[11] M. Mohammadi and B. Vahidi, Model of corona for transient study,presented at the IEEE Region TENCON, Melbourne, Qld., Australia,2005.

[12] R. Davis, R. W. E. Cook, and W. G. Standring, The surge coronadischarge, Proc. Inst. Elect. Eng. Monographs, vol. 108, pp. 230239,1961.

[13] P. S. Maruvada, H. Menemenlis, and R. Malewski, Corona charac-teristics of conductor bundles under impulse voltages, IEEE Trans.Power App. Syst., vol. PAS-96, no. 1, pt. 1, pp. 102115, Jan. 1977.

[14] C. F. Wagner, I. W. Gross, and B. L. Lloyd, High-voltage impulsetests on transmission lines, Power App. Syst., Part III. Trans. Amer.Inst. Elect. Eng., vol. 73, pp. 196210, 1954.

YANG et al.: LIGHTNING IMPULSE CORONA CHARACTERISTIC OF 1000-kV UHV TRANSMISSION LINES 2525

[15] C. F. Wagner and B. L. Lloyd, Effects of corona on traveling waves,Power App. Syst., Part III. Trans. Amer. Inst. Elect. Eng., vol. 74, pp.858872, 1955.

[16] M. Ouyang and P. G. Kendall, Tests on distortion and attenuationof waves on an overhead line, IEEE Trans. Power App. Syst., vol.PAS-94, no. 2, pt. 1, pp. 498507, Mar. 1975.

[17] C. Gary and D. Cristescu, Distortion and attenuation of travellingwaves caused by transient corona, Electra, vol. 131, pp. 170183,1990.

[18] G. V. Podporkin and A. D. Sivaev, Lightning impulse corona charac-teristics of conductors and bundles, IEEE Trans. Power Del., vol. 12,no. 4, pp. 18421847, Oct. 1997.

[19] T. Narita and S. Okabe, Propagation characteristics of lightningsurge with corona discharge on UHV designed transmission line, (inJapanese) Trans. Inst. Elec. Eng. Jpn., vol. 122-B, no. 2, pp. 307313,20022.

[20] J. G. Anderson, Transmission Line Reference Book, 345 kV and Above,2nd ed. Palo Alto, CA, USA: Elect. Power Res. Inst., 1982, ch. 12.

[21] G. Carrara, L. Dellera, and G. Sartorio, Switching surges withverylong fronts (above 1500 s): Effect of front shape on dischargevoltage, IEEE Trans. Power App. Syst., vol. PAS-89, no. 3, pp.453456, Mar. 1970.

[22] C. Menemenlis and K. Isaksson, Influence of the various parts ofthe switching impulse front on discharge development, IEEE Trans.Power App. Syst., vol. PAS-94, no. 5, pp. 17251733, Sep./Oct. 1975.

[23] J. R. Marti, F. Castellanos, and N. Santiago, Wide-band corona circuitmodel for transient simulations, IEEE Trans. Power Syst., vol. 10, no.2, pp. 10031013, Apr. 1995.

[24] K. C. Lee, Non-linear corona models in an electromagnetic transientsprogram (EMTP), IEEE Trans. Power App. Syst., vol. PAS-102, no.9, pp. 1293612942, Sep. 1983.

[25] S. J. Carneiro and J. R. Marti, Evaluation of corona and line models inelectromagnetic transientssimulations, IEEE Trans. Power Del., vol.6, no. 1, pp. 334342, Jan. 1991.

[26] H. Motoyama and A. Ametani, Development of a linear model forcorona wave deformation and its effect on lightning surges, Elect.Eng. Jpn., vol. 107, no. 2, pp. 98106, 1987.

Pencheng Yang, photograph and biography not available at the time ofpublication.

Shuiming Chen (SM02), photograph and biography not available at the timeof publication.

Jinliang He (F08) was born in Changsha, China,in 1966. He received the B.Sc. degree in electricalengineering from Wuhan University of Hydraulicand Electrical Engineering, Wuhan, China, in 1988,the M.Sc. degree in electrical engineering fromChongqing University, Chongqing, China, in 1991,and the Ph.D. degree in electrical engineering fromTsinghua University, Beijing, China.He became a Lecturer in 1994, and an Associate

professor in 1996 in the Department of Electrical En-gineering, Tsinghua University. From 1994 to 1997,

he was the Head of High Voltage Laboratory at Tsinghua University. From 1997to 1998, he was a Visiting Scientist at the Korea Electrotechnology ResearchInstitute, Changwon, Korea, involved in research on metaloxide varistors andhigh-voltage polymeric metaloxide surge arresters. In 2001, he was promotedto Professor at Tsinghua University. Currently, he is the Vice Chief of the HighVoltage Research Institute at Tsinghua University. His research interests includeovervoltages and electromagnetic compatibility in power systems and electronicsystems, lightning protection, grounding technology, power apparatus, and di-electric material. He is the author of five books and many technical papers.