S

Aa

b

a

AA

KLCGDHE

1

pftavp

acdlablp(aiig(

r

0d

Electric Power Systems Research 79 (2009) 428–435

Contents lists available at ScienceDirect

Electric Power Systems Research

journa l homepage: www.e lsev ier .com/ locate /epsr

urge voltages and currents into a customer due to nearby lightning

kihiro Ametania,∗, Kae Matsuokaa, Hiroshi Omurab, Yoshiyuki Nagaib

Power System Analysis Laboratory, Doshisha University, Kyo-tanabe, Kyoto 610-0321, JapanPower Engineering R and D Center, Kansai Electric Power Co., Amagasaki, Osaka 661-0794, Japan

r t i c l e i n f o

rticle history:vailable online 18 October 2008

a b s t r a c t

This paper presents experimental results of lightning surges incoming into a customer due to lightningto an antenna of the customer, a pole and a ground nearby the customer, and briefly discusses lightning

eywords:ightning surgeustomerroundingistribution line

current distribution in the customer, a distribution line and a telephone line. Based on experimentalresults, modeling of each component is explained, and EMTP simulations are carried out. The groundvoltage rise is represented by a mutual resistance between grounding electrodes. EMTP simulation resultshave been observed to agree qualitatively with the measured results, and it becomes possible to investigatelightning surges and current distribution in a customer by an EMTP simulation.

dsa

2

2

ohiami

(

(

ome applianceMTP

. Introduction

Lightning overvoltages are important not only from the view-oint of insulation coordination of transmission systems but alsorom the viewpoint of protection of customers’ equipments andelecommunication systems. A number of disturbances of homeppliances due to the lightning have been informed [1–5]. In a sur-ey of the disturbances in a utility in Japan, nearly 100 disturbanceser year are found to be caused by the lightning.

This paper investigates experimentally an incoming path oflightning surge into a customer due to lightning nearby the

ustomer [6,7]. There exist four incoming paths: (1) low-voltageistribution line (feeder) through a distribution pole, (2) telephone

ine, (3) customer’s TV antenna and (4) grounding electrode ofcustomer’s electrical equipment. A lightning strike to a distri-

ution line and an induced lightning voltage to the distributionine result in the path (1). Similarly a lightning strike to a tele-hone line and an induced lightning voltage result in the path2), and the same is applied to the path (3). A lightning strike toground, a wood or a distribution pole nearby a customer results

n a ground potential rise (GPR) due to a lightning current flow-ng into the ground [8,9]. The current causes a potential rise to a

rounding electrode of a customer’s equipment which is the path4).

Based on the measured results, modeling of electrical elementselated to the above paths for a lightning surge simulations are

∗ Corresponding author.E-mail address: aametani@mail.doshisha.ac.jp (A. Ametani).

cbahtci(

378-7796/$ – see front matter © 2008 Elsevier B.V. All rights reserved.oi:10.1016/j.epsr.2008.09.005

© 2008 Elsevier B.V. All rights reserved.

eveloped, and EMTP simulations are carried out by the model. Theimulation results are compared with the measured results and theccuracy of the modeling method is discussed.

. Experiment of incoming surge into house

.1. Experimental conditions

Kansai Electric Power Co. (KEPCO) has carried out a numberf experiments to investigate lightning currents flowing into aouse [6,7]. An impulse current representing a lightning current

s injected into an antenna, a distribution pole and a ground fromn impulse current generator (IG, maximum voltage 3 MV, maxi-um current 40 kA). Fig. 1 illustrates an experimental setup. An

mpulse current from 500 to 5000 A is applied to:

(a) an antenna: Fig. 1(A);b) a distribution pole: Fig. 1(B);

(c) a structure nearby a house: Fig. 1(C);d) a messenger wire of a telephone line: Fig. 1(D).

In a test-yard of the KEPCO, 3 distribution poles and 2 telephoneompany (NTT) poles are constructed and 6600/220/110 V distri-ution lines with a pole transformer are installed as in Fig. 1. Also,telephone line and the messenger wire are installed. A model

ouse is built nearby the pole transformer, and a feeder line fromhe transformer is led in. In the model house, model circuits of an aironditioner and a fax machine are installed as shown in Fig. 2. Thosenvolve a surge protective device SPD, a kind of a surge arrestorair conditioner: operating voltage 2670 V, fax: 2800 V, NTT SPD:

A. Ametani et al. / Electric Power System

Fg

5ti

ot

2

ctf

Fv

2

tItahtaFi1fltl

2

8ndl

2

(

Fout to a distribution line. The current into the NTT-SPD protectorflows out through the telephone line. No current flows into the fax,because the operating voltage 2800 V of a surge arrestor withinthe fax is higher than the voltage across the fax, i.e., the voltage

ig. 1. Lightning and its path to a house. Lightning to: (A) antenna, (B) pole, (C)round and (D) telephone (messenger) line.

00 V). The poles, the transformer, the telephone line (NTT) andhe air conditioner (home appliance) have their own grounding asn Fig. 2.

Table 1 summarizes the test conditions and measured resultsf maximum voltages and currents through the grounding resis-ances.

.2. Experimental results

Fig. 2 illustrates the experimental circuit for various conditionsorresponding to A–C in Fig. 1. An impulse current is applied to (a)he antenna, (b) the top of the distribution pole and (c) a groundrom the impulse current generator IG.

Fig. 3 shows experimental results for the conditions inig. 2(a)–(c). From the experimental results, the following obser-ations have been made.

.2.1. Lightning to antennaWhen lightning strikes the antenna of a house in Fig. 2(a),

he lightning current flows out (1) to a distribution line I1 (In,t) through feeders in the house, and (2) to the earth I2 (Ia, Inp)hrough grounding electrodes of home appliances. The ratio of I1nd I2 is dependent on the grounding resistances (1) seen from theouse feeder to the distribution line, (2) seen from the house tohe telephone line and (3) of the grounding electrodes of the homeppliances. In the experimental results one of which is given inig. 3(a) with the applied current I0 = 739 A, about 85% of the currentnto the antenna flows out to the distribution line. The remaining5% is estimated to flow into the other houses. A large current 543 Aowing into the distribution pole is caused by a flashover betweenhe transformer grounding lead and the steel of the pole and theow grounding resistance of the pole.

.2.2. Lightning to distribution poleFig. 3(b) shows a measured result when I0 is 2416 A. In this case,

.1% of I0 flows into the grounding resistances Rt of the transformereutral, 8.5% flows into the grounding resistance Ra of the air con-itioner, and 2.8% to the grounding resistance Rnp of the telephone

ine SPD.

.2.3. Lightning to groundingA part of the applied current I0 flows into the air conditioner

Ia max = 58.2 A for I0 = 2815 A) and the NTT SPD (Inp max = 56.6 A) in

s Research 79 (2009) 428–435 429

ig. 3(c), case C1. The current flowing into the air conditioner flows

Fig. 2. Experimental circuit and lightning current path.

430 A. Ametani et al. / Electric Power Systems Research 79 (2009) 428–435

Table 1Test conditions and results

Case

A B C1 C2

306a 739a 1247a 1660a 602a 1183a 1977a 2416a 676a 1532a 2815a 4134a 2771a

Grounding resistance (�)Rp 17 17 17 17 19 19 19 19 19 19 19 19 19Rt 89 89 89 89 75 75 75 75 85 85 85 85 85Ra 830 830 830 830 100 100 100 100 140 140 140 140 –Rnp 900 900 900 900 300 300 300 300 380 380 380 380 380Rn1 153 153 153 153 150 150 150 150 100 100 100 100 100

Voltage (kV)Vo 11.6 25.9 41.5 56.1 10 19.3 31.5 37.6 55.4 88 122 146.3 122Vt 3.8 7.7 12.9 17.5 3.8 5.6 21.2 24.6 – – 2.5 – 2.5Va – – – – 0.8 5.1 19.6 22.8 1.8 2.2 3.1 4.4 3.1Vnp – – – – 1.3 2.9 16.7 19.1 1.3 2.9 5 6.3 0.8Vp 3.5 7.5 12.2 16 – – – – – – – – –

Current (A)Ia – – – – 0 44 177 206 0 27.5 58.2 83.1 55.7Inp – – – – 3.6 5.9 52 68 13.2 31.2 56.6 80.2 –It 34.4 72 120 159 3.4 −27 174 197 – – – – –

dl

r

3

3

hv[lmpTtifsds[

firof

Z

wp

gw6

befPlTZ

fNtaiflF

ttgbr2Imsbahn1

rat

Ip 247 500 888 1150 – –If – – – – – –

a Applied I0 A.

ifference between the feeder line in the house and the telephoneine.

Compared with case C2, no air conditioner grounding, the cur-ents flowing into the house increases heavily.

. Modeling and simulation of incoming surge

.1. Modeling

The distribution line, the pole and home appliances in theouse in Fig. 2 can be readily represented by horizontal andertical distributed line models and lumped-parameter circuits10,11]. The grounding electrodes of the pole, the telephoneine SPD and the home appliances if those are grounded are

odeled by a combination of a distributed line and a lumped-arameter circuit to simulate the transient characteristic [12].his paper, however, adopts a simple resistance model of whichhe resistance value is available from the experiments explainedn Refs. [12,13], because the vertical grounding electrode usedor a home appliance is short and the transient period is farmaller in the phenomenon investigated in this paper. A protectiveevice installed in a home appliance and the NTT SPD are repre-ented simply by a time-controlled switch prepared in the EMTP14].

A model circuit for an EMTP simulation is shown in Fig. 4. In thegure, Zp is the surge impedance of a distribution pole which is rep-esented by a lossless distributed line with the propagation velocityf 300 m/�s. The surge impedance is evaluated by the followingormula of a vertical conductor [10]:

p [˝] = 60

{ln

(hp

rp

)− 1

}(1)

here hp is the height of the pole and rp is the radius of the

ole

Rp = 17–19 � is the grounding resistance of the pole. The value isiven by a utility [6,7]. Tr is a pole mounted distribution transformerhich is represented by an ideal transformer with the voltage ratio600:110 V and stray capacitances.

3

s

– – – – – – –– – 0 0 0 0 31.8

TRN-TRNB is a grounding lead of the transformer representedy a lossless distributed line of which the surge impedance isxplained in Ref. [15]. Rt is the grounding resistance of the trans-ormer grounding lead and is given as in Table 1 by a utility [6,7].W is a phase wire of a distribution line and is modeled by aossless distributed line with the surge impedance Zpw = 530 �.he other end of PW1 is terminated by the matching impedancepw.

TRAL-FAX and -HOME1 are a feeder line from the pole trans-ormer to a house. For the transformer, A is for phase A and

for neutral. The steady-state AC voltage between the phase Ao the neutral is 114 V in Japan. In the case of lightning to anntenna, case A, a flashover between the transformer ground-ng lead and the steel of a distribution pole was observed. Theashover is modeled by short circuiting those by resistance Rp inig. 4(a).

The feeder line is represented by a lossless distributed line withhe surge impedance of 560 � and the velocity of 300 m/�s. Ra ishe grounding resistance of an air conditioner of which the value isiven in Table 1 [6,7]. Za represents the air conditioner expressedy a lead wire inductance 1 �H and by a time-controlled switchepresenting a surge arrester operating when the voltage exceeds670 V in parallel with the resistance. Zb represents a fax machine.t is expressed in the same manner as Za except that the other ter-

inal of Zb is connected to a telephone line through an NTT SPD, aurge arrester valve with the operating voltage of 500 V representedy a time-controlled switch. The telephone line is represented bylossless distributed line with the surge impedance of 560 � ande propagation velocity of 200 m/�s and the other end is termi-ated by the matching resistance with the grounding resistance of53 �.

IG is an impulse current source used in the experiment and isepresented by a charged capacitance and a resistance [6,7]. Rm ismutual grounding impedance between various grounding elec-

rodes, and the value of 2–10 � is adopted [9,13].

.2. Simulation results

Table 2 summarizes simulation conditions and results, and Fig. 5hows simulated voltage and current waveforms corresponding to

A. Ametani et al. / Electric Power Systems Research 79 (2009) 428–435 431

of tran

tpm

(

Fig. 3. Experimental results

he measured results in Fig. 3. From the simulation results in com-

arison with the measured results, the following observations areade.

1) In Fig. 5(a) when lightning strikes the antenna, simulationresults of transient voltages and currents agree well with the

(

sient currents and voltages.

measured results in Fig. 3(a) except the wavefront of the applied

current and voltage at the top of the pole and the voltage at theprimary winding on the transformer.

2) In the case of lightning to the pole top in Fig. 5(b), voltagedifferences between the top and the bottom of the pole andat the air conditioner grounding electrode are observed to be

432 A. Ametani et al. / Electric Power Systems Research 79 (2009) 428–435

Fig. 4. Model circuits for EMTP simulations: (PW) power line and (CW) communication (NTT) line.

A. Ametani et al. / Electric Power Systems Research 79 (2009) 428–435 433

ulatio

(

Fig. 5. Sim

smaller than those of the measured results in Fig. 3(b). Other-

wise the simulation results show a reasonable agreement withthe measured results.

3) When lightning hits ground nearby a house, the simulationresults in Fig. 5(c) shows a qualitatively satisfactory agreementwith the measured results in Fig. 3(c), and thus the proposed

lr

n results.

approach to deal with a ground voltage rise by means of a

mutual impedance is confirmed to be adequate.

It, however, requires a further improvement of the overall simu-ation method to achieve a quantitative agreement with a measuredesult. For example, a surge protecting device has to be carefully

434 A. Ametani et al. / Electric Power Systems Research 79 (2009) 428–435

Table 2Simulation results corresponding to Table 1

Case

A B C1 C2

306a 739a 1247a 1660a 602a 1183a 1977a 2416a 676a 1532a 2815a 4134a 2771a

Voltage (kV)Vo 87.2 174.3 261.4 348.5 16.2 30.6 49.6 60 50.8 81.24 121.9 162.4 121.9Vt 17.6 35.2 52.8 70.5 13.6 25.6 41.5 50.2 1.7 2.7 4 5.3 3.4Va – – – – 5.9 11.2 18.2 22 4.3 6.9 103.3 137.8 103.8Vnp – – – – 9.8 18.4 29.9 36.2 2.9 4.6 6.9 9.3 6.5Vp 22.4 44.8 67.1 89.5 – – – – – – – – –

Current (A)Ia – – – – 61 114.9 186.5 225.6 −19.8 −31.3 −47.6 −62.6 −53.9Inp – – – – 23.7 44.7 72.5 87.8 −8 −12.7 −19 −25.4 –It 198 395 593 791 143 269.3 437.2 528.8 – – – – –Ip 240 480 721 961 189.4 356.7 579.2 700.5 – – – – –

– – −2.4 −3.9 −6 −7.9 7.3

rAtt

4

rnwicmudttatitr

flebtg

A

A

F

V

W

I

R

Z

R

R

If – – – – – –

a Applied I0 A.

epresented based on its circuit and the nonlinear characteristic.lso, a grounding impedance should be modeled considering its

ransient characteristic rather than a simple resistance adopted inhis paper.

. Conclusions

This paper has shown experimental and EMTP simulationesults of a lightning surge incoming into a house due to light-ing nearby the house. The simulation results agree qualitativelyith the experimental results, and thus the simulation models

n the paper are said to be adequate. The measurements beingarried out in different time periods for 3 years, the experi-ental conditions such as the earth resistivity, a voltage probe

sed and a reference voltage line for each measurement areifferent, and, the difference is not considered in the simula-ions Some oscillations observed in the measured the estimatedo be caused by mutual coupling between the measuring wiresnd feeder lines in the experiments. Also, a grounding elec-rode may have coupling to the other electrodes. A furthermprovement for modeling the experimental circuits is requiredo achieve a higher accuracy in comparison with the measuredesults.

Lightning surge voltages in a customer and lightning currentsowing out to a distribution line can be made clear based on thexperimental and simulation results. The results are expected toe applied to protection coordination of SPDs for customers and aelephone line, and to investigate the necessary of home appliancerounding.

ppendix A. Appendix A

.1. Steady-state voltages in Fig. 5(b)

At around 5 �s, Fig. 5(b) is rewritten as Fig. A.1(a). It is clear inig. A.1(a) that the voltages Va and Vt are the same:

a [kV] = Vt = E0 − R0I1 = 16.4

hen calculating currents, Fig. A.1(b) is used:

1 = 932 A, I2 = 37 A, I3 = 169 A

0I1 + Z1(I1 − I2) + Rm(I1 − I3) = E

Fig. A.1. Equivalent circuit of Fig. 5(b) at around 5 �s.

1(I2 − I1) + RnI2 + Rnp(I2 − I3) = 0

tI3 + Rm(I3 − I1) + Rnp(I3 − I2) = 0

eferences

[1] B. Smith, B. Standler, The effects of surge on electronic appliances, IEEE Trans.,PWRD-7 3 (1992) 1275.

[2] M. Kawahito, Investigation of lightning overvoltages within a house by means ofan artificial lightning experiment, in: R&D News Kansai Electric Power, Septem-ber, 2001, pp. 32–33.

[3] Y. Imai, N. Fujiwara, H. Yokoyama, T. Shimomura, K. Yamaoka, S. Ishibe, Anal-ysis of lightning overvoltages on low voltage power distribution lines due todirect lightning hits to overhead ground wire, IEE Jpn. Trans. PE 113-B (1993)881–888.

[4] T. Hosokawa, S. Yokoyama, T. Yokota, Study of damages on home electricappliances due to lightning, IEE Jpn. Trans. PE 125-B (February (2)) (2005)221–226.

[5] Y. Nagai, H. Sato, Lightning surge propagation and lightning damage risk acrosselectric power and communication system in residential house, IEICE Japan,Research Meeting (2005), EMC-05-18.

System

[

[

[

[

[[

Yoshiyuki Nagai was born in March 1949. He graduatedfrom Osaka Prefectual Collage in March 1969, and joinedKansai Electric Power Co. in April 1969. He has beenengaged in Power Engineering R&D center of KEPCO, espe-cially in the field of grounding and lightning.

A. Ametani et al. / Electric Power

[6] Y. Nagai, N. Fukusono, Lightning surge propagation on an electric power facil-ity connected with feeder lines from a pole transformer, in: KEPCO ResearchCommittee of Insulation Condition Technologies, June, 2004.

[7] Y. Nagai, Lightning surge propagation into a model house from various places,in: KEPCO Research Committee of Insulation Condition Technologies, March,2005.

[8] S. Yokoyama, H. Taniguchi, The third cause of lightning faults on distributionlines, IEE Jpn. Trans. PE 117-B (October (10)) (1997) 1332–1335.

[9] A. Ametani, K. Hashimoto, N. Nagaoka, H. Omura, Y. Nagai, Modeling of incominglightning surges into a house in a low-voltage distribution system, in: Proceed-ings of the EEUG 2005, Warsaw, Poland, September, 2005.

10] A. Ametani, K. Kasai, J. Sawada, A. Mochizuki, T. Yamada, Frequency-dependentimpedance of vertical conductors and a multiconductor towermodel, IEE Proc.GTD 141 (July (4)) (1994) 339–345.

11] A. Ametani, K. Koji, Y. Yuki, N. Mori, A frequency characteristic of the impedanceof a home appliance and its equivalent circuit, IEE Japan, Annual Conference,March, 1994, No. 1405.

12] D. Soyama, Y. Ishibashi, N. Nagaoka, A. Ametani, Modeling of a buried conductorfor an electromagnetic transient simulation, in: Proceedings of the ICEE2005,Kunming, China, July, 2005 (SM1-04).

13] M. Nayel, A study on transient characteristics of electric grounding systems,Ph.D. Thesis, Doshisha University, November, 2003.

14] W. Scott Meyer, EMTP Rule Book, B.P.A., Portland, OR, 1982.15] T. Mozumi, T. Ikeuchi, N. Fukuda, A. Ametani, S. Sekioka, Experimental formulas

of surge impedance for grounding lead conductors in distribution lines, IEE Jpn.Trans. PE 122-B (February (2)) (2002) 223–231.

Akihiro Ametani was born on 14 February 1944. He receivedPh.D. degree from UMIST, Manchester in 1973. He waswith the UMIST from 1971 to 1974, and Bonneville PowerAdministration to develop EMTP for summers from 1976to 1981. He has been a professor at Doshisha University

since 1985 and was a professor at the Chatholic Univer-sity of Leaven, Belgium in 1988. He was the Director of theInstitute of Science and Engineering from 1996 to 1998,and Dean of Library and Computer/Information Center inDoshisha University from 1998 to 2001. Dr. Ametani is aChartered Engineer in UK, a Fellow of IEE and IEEE.

s Research 79 (2009) 428–435 435

Kae Matsuoka was born in October 1984. She graduatedfrom Electrical Eng. Dept. of Doshisha University, Kyoto,Japan in March 2007 and is an M.Eng. student in the Grad-uate School of the same university.

Hiroshi Omura was born in 1960. He graduated from Mat-sue Technical High School, Shimane Prefecture and joinedKansai Electric Power Co., Inc. (KEPCO), Osaka, Japan in1979. He spent 2 years for study and research in thedepartment of electrical engineering in Kyoto Universityfrom April 1985 to March 1987. He has been engagedin Power Engineering R&D Center of KEPCO since July2003.