An Analysis Of Switching Overvoltages in the EGAT 500 kV Transinission System

K. Ngamsanroaj, and W. Tayati: PhD

Ahsrrocf--The level of transient overvoltage has a strong effect to power system reliability. I t has already been k n o w n that switching o\w-voltages are the most inipoflant criteria which the design of insulation of extra high voltage (EHV) systems is based on. The rrduction of switching surges is tlierefore an economic necessity. This paper illustrates switching overvoltages resulted from various switrhing conditions applied to all the existing 500 k\' transmission lines of t h e Electricity Generating Authority of Thailand (EGAT) system. The study is performed with computer transient simulations using PSCAD/EMTDC program. Details of the system modeling are outlined. The appropriate representation for the various components such as transmission lines, transiorniers, equivalent sources, shunt reactors, surge arresters a n d circuit breakers has been chosen. Switching representing t w o niost severe causes: - line energization and line re-cviergization due to faults i.e. single phase to ground and three phases to ground are considered. Results are analyzed in this paper.
Irtrfex 7-er??z.c-- ER1TP, 1 ransmission Line Modeling.

S w it r h in g

Overvo It apes

a nd

1. INTRODUCTIOK N order to design and select the proper insulation level for transmission systems, it is necessary to analyze the different kinds of overvoltages, especially, the transient overvoltage that is a short phenomenon of power system initiated by action o f circuit breakers and switches by faults. These actions include energization. re-energization, reclosing and fault clearing. The level of transient overvoltage has a strong effect lo power system. As a result, 500 kV EHV. overvoltage from switching is the most important condition to consider in this regard, there is incentive to minimize the overvoltage to lower new equipment costs or limit the voltage stresses on aging equipment still in service. The range of frequencies of primary interest in a switching transients study varies from the fundamental power frequency up to about 10 kHz. Therefore the proper representation must be chosen for the various components such as transmission lines and cables. transformers, source equivalents, loads and circuit breakers. Equipment modeling aspects for the analysis of switching overvoltages are the principal subject of this paper I 11.

K Npamsanroai is with Department of Electrical knpineering Chiang Mal Universitv Chianp Mal 50100 Thailand (e-mail tpckc@epat or th', W 1ayati. PhD i s with the Department of Electrical Engineering. Chianp 50100 Thailand re-mail Mal Universi~ Chiang Mal worawit@eng cmu ac th)

Many simulation approaches have been introduced for voltage transient simulation. The parameters of transmission lines with ground return are highly dependent on the frequency (21. Accurate modeling of this frequent! dependence over the entire frequency range of the signals is of essential imponance for the correct simulation of electromagnetic transient conditions. In transient simulations. the frequency-dependent representation of transmission lines required only 10-30% more computer time than the constantparameter simulation. In this [3], closing resistors may be eliminated from EHV power circuit breakers when pole closing or polarity closing is used on transmission lines that have meta oxide surge arresters only at terminals, This paper reports the results of computer studies and field tests that validate the concept and techniques for statistical studies of specific cases. The investigation concludes that 500 kV lines up to 320 km may be energized successfully without the use of closing resistors. Working Group 3.4.1 1 on Surge Arrester Modeling has been reviewed a number of ways to model metal oxide artesters.[4] Lab test data of metal oxide arrester discharge voltage and currents available to the working group have indicated that metal oxide arresters have dynamic characteristics that are significant for studies involving lightning and other fast-front surges. A model is described which will give an appropriate voltage response for a current surge which has a time-to-crest anywhere in the range of 0.5 ps to 45 ps. The work presents a method for generating the parameters of the model from published manufacturer's data. For switching surge studies, an arrester can be modeled by choosing an appropriate nonlinear V-1 characteristic. There is no need for a frequency-dependent model. Bonneville Power Administration's first SO0 kV back-to-back shunt capacitor installations were described in [SI.The primary purpose of the capacitor banks is to support A C system voltage for transient stability following an outage of the HVDC intertie line. Design studies and equipment requirements are detailed herein in addition to EMTP studies and subsequent field test results. Reference [6] described the study made to assess the electrical stresses imposed on line circuit breakers of a new SO0 kV transmission system. Digital simulations have been made to calculate the switching transients appearing as a result of line energization and reclosing, clearing of terminal and short-line faults. no-load line opening in normal and faulted conditions. and out-of phase switching. Two most severe causes of switching overvoltages: - line energization and line re-energization due to faults i.e. single
14C

0-7803-7863-6/03/$17.00 02003 IEEE

Authorized licensed use limited to: UNIVERSIDADE ESTADUAL DE CAMPINAS. Downloaded on June 30,2010 at 14:59:33 UTC from IEEE Xplore. Restrictions apply.

phase to ground and three phases to ground will be investigated in this paper for the entire existing EGAT 500 k V network. The EMTI' program is used to perform the siniulation, as it proved to be the highest and most accurate digital simulation environment for handling such a problem

piZFk?T\-

j l

11. DESCRIPTION T H E SYSTEM OF

At the early stage of EHV transniission line construction in 1982, EGAT had employed General Electric (USA) to stud! 500 kV transmission systems? also to cooperate with consultant to study the highest stress of switchinp overvoltages and impose Switching Impulse Withstand Voltage Level (SI WL). 1-ransient Network Analyzer ( T N A ) had been used for analyzing process. This is used to analyze transient condition of the specific system. design the first stage 500 kV transniission lines (Mae Moh - Tha TaKo Circuit 1)- and report TNA study on the various case. However? when E G A l wants to expand EHV transmission systems which need additional analysis and design, TNA will not be a proper tool. The reasons are that its components and circuit implementation consume time and budget. Therefore. the consecutive transmission lines construction prqiect (from 500 kV Mae Moh - Tha TaKo circuit 2 to present) ha5 selected Electromagnetic 1-ransient Program (EMTP) which is powerful and more flexible to help the design process for each construction phase. At present: EGAT has expanded to total about 2,600 circuits -km of 500 kV (fig. 1). However, the overvoltage estimation of overall system which may occur and reach the withstand voltage has not yet been conducted. EGAT requires analyzing the voltage level of switching overvoltage of existing 500 kV transmission lines.

A
Pluak Daen

I I/
I

; Ratchaburi k--. _ Fig 1 Existing EGAT 500 kV Iransmission Sysieni

TABLE 1
CHARACTERISTIC OF THE LINE5

111. MODELING REQUIREMENTS

A.

Transmission Lines

For each transmission line represented in the network, dimensions and data are required as depicted in table 1. This can be given at the tower, and include conductor sag. Shield wire dimensions and resistance are also provided. The transmission line data required includes: transniission line conductor diameter and resistance per unit length: total length of transmission line, phase transforniation data and distances between phase transformation, spacing between conductors in a phase bundle, spacing between phases, shield wire diameter and resistance per unit length, height of each conductor and shield wire at the tower and sag to midspan, tower dimensions: and ground conductivity. The selected transmission line models are distributed parameter models based on the traveling time and characteristic impedance of the line.

150

Authorized licensed use limited to: UNIVERSIDADE ESTADUAL DE CAMPINAS. Downloaded on June 30,2010 at 14:59:33 UTC from IEEE Xplore. Restrictions apply.

are listed in table 2

B.

Trutisjot-met

For each transformer to be represented in detail with thi5 existing information : MVA rating. winding configuration and voltage, tap change ranges and normal setting. leakage reactance between windings. knee point of transformer core saturation characteristic in per unit of rated flux or voltage. and estimated saturated air core reactance.

TABLE Z
SWITCHmC OVERVOLTAGES OBIAINED FOR SYSTEM SIMULATIOK B Y LINE ENERCIZATIOK

C. Circiiii BreaketThe locations of the circuit breakers that will be switched are identified on the study system. Other parameters o f the circuit breakers are determined: protection delay or clearin? times, maximum fundamental frequency switching voltage. maximum capacitive switching capabilic. reclosinp sequences, rated transient recovery voltage and maximum rate of rise of transient recovery voltage. mechanical closinp time and variation in pole closing times, and closing resister.

D. Surge Arresters
The installed location and rating of surge arresters are provided. The niax~mtiniratings. and in particular the energ? absorption capability will be determined with the studies Considering the above table, the highest maximum switching overvoltage depicted is the simulation 0 1 energization from Mae Moh to Tha Tako. The maximum overvoltages for line energization have slight constan1 differences at each location. Out of the aforementioned simulation, it can be concluded that some overvoltages are very severe, and means of reduction should be used on all the lines. Breakers equipped with surge arrestors are to be installed. Faults with the most severe overvoltages are re-simulated on the PSCADEMTDC. The case of single phase and three phases faults with line re-energization are considered at the location subject to the highest switching overvoltage. In the series of simulations, the maximum overvoltages obtained during different line switching operations are determined. Line energization and re-energization operations are examined. Each line in the system is connected to two series time controlled switches at its two ends. The switched close simultaneously to model the line energization and/or line reclosing, according to the status of the system fauli conditions. Line energization is made and remains in the network for the period of transient analysis. Single phase to ground fault is applied at Mae Moh at t = 0.02 s. cleared at ~ 0 . 0 5 and re-energized at ~ 0 . s at no load condition. The s 1 result is presented in fig.2. The reduction of voltage level is made by shunt reactors connected at both ends as shown in fig. 3.

E. Sliunt Reucrors
The location of shunt reactors is identified. This will included whether they are line connected or bus connected. The reactor rating is supplied along with its characteristic if it is a saturating reactor.

1v. STUDIES AND RESULTS

In the series of simulations at no load, bus faults at line side are applied at different locations on the system and the transient overvoltages are recorded in the system, with and without fault clearing. Three phases to p o u n d and single phase to ground faults are applied on a 3 phase bus at various times and determining the worst overvoltage. Fault clearing is usually accompanied with line opening. where the circuit breakers open at both ends on the line where the fault occurs to isolate the fault. Especially for no load condition the effect is the most severe. From the existing 500 kV system, surger arresters and shunt reactors are included in the model. The circuit breaker closes at one end with the other end open for the weak system condition when energizing. When line connected surge arresters are used, the overvoltages are reduced to lower level. T h e switching overvoltages obtained on the various lines and during the various conditions of system are listed below. The voltages are the niaxinium recorded with each case havinp 300 statistically controlled closings. The switching overvoltages obtained for system operation

151

Authorized licensed use limited to: UNIVERSIDADE ESTADUAL DE CAMPINAS. Downloaded on June 30,2010 at 14:59:33 UTC from IEEE Xplore. Restrictions apply.

Re-energizing due to slngle phases to ground fault

./ ,h

V. C O N C L U S I O ~

PnB\

Ac , -2

5
m
,D

t12-02L

jb

,6!J

r/ \

p-.^-.-- /\,, 1 (p /
;

\/

1;

fL\,

Pt,
1

ii/

System switching operations are not always plannetj events. Some disturbance may cause a switching operation during a period of very light load. The response of a verr lightly loaded system I S nearly the same as the response oi an unloaded system. Therefore. for ease and simplicity. switchinp represented on the system. Switching operations on an unloaded system will produce the most severe transierit overvoltages. Energizing and reclosing transient overvoltages are investigated statistically - that is, for each system condition investigated, 300 switching operations are made. each statistically controlled with respect to the source bus driving voltage and with the three breaker poles closing randoml,) with respect to each other, but within a specific closing span guaranteed by the circuit breaker manufacturer. With the studies carry out it was possible to find the switchirig overvoltapes that affect the existing 500 kV transniission lines of EGA7 system. Analyzing the obtained results. it can be concluded the most severe situations for the system. It is possible with these results to proceed With the selection of the protections in order to minimize the impact of the switching overvoltages in the network. The results could be considered as a reference for operation, insulation coordination improvement and further construction.

Time (sec)

I ip 2 Single phase to ground fault and fault clearing at Mae Moh Tha Taho
-

Remne~gigng to single phase to ground faujt due

With Shunt Reactors at both ends

Time (sec)
Fig 3 Single phase to ground fault and fault clearinp at Mae Moh Tha 1-aho with Shunt Reactors at hoth ends

Fig. 4 depicts the three phases fault clearinp with line reenergization at no load condition. I t is found that the highest maximum overvoltage when energize from Mae Moh to Tha Tako is about 3.24 pu. without any protective system such as shunt reactor or surge arrester. In fig. 5 . Shunt reactors connect at both ends reduces the amount of maximum overvoltage to 2.63 pu.
Rp,enelgizlng due to thres phases to ground fauh
M,

VI. ACKNOWLEDGMENI
This work would not have been possible without the contribution of t h e company EGAT. in panicular the 1 ransmission System Department. The authors are very grateful. in particular to Mr. Luechai Surapongohan and Dr. Sulhep Chimklai for their technical suppor~.

+PW\

PnB \

G PhC \

REFERENCES
-3

0 03: 0 013 0 0% 4 15.08.09. Modeling and 0 12f 0 1 6 [ 11 IEEE PES Working Group
Time (sec)

_.

Analysis of System 1 ransients Using Digital Programs. IEEEE

PES Special Publication, pp. 4-1 - 4-24. 1998.

Fig 4 Threc phases 10 ground fault and fault cleartnp at Mae Moh Tha Tako

Re-energizingdue

ground faujt-

With Shunt Reactors at both ends

E PhA V 47,

- PhB V

z- V PhC

123 Marti J.R.. Accurate Modeling of Frequency-Dependent 1-ransniissioii Lines in Electromagnetic 1-ransient Simulations. IEEE Trunsaction on Power Appuruius and Svsiems, vol. PAS-I0I.No. 1, pp. 147-155. January 1982.

[3] Legate A.C.. Brunke J.H.. Ray J.J. and Yasuda E.J.. Elimination of closing resistor on EHV circuit breakers. 1EEE irunsaciion on Power Deliver)). vol. 3. no. 1, pp. 223-231. January 1988

Time (sec)
Fig 5 Three phases to ground fault and fault clearine at Mae Moh Tha Tako with Shunt Reactors at both ends

[4] IEEE Working Group 3.4.1 1, Modeling of Metal Oxide ~ ~ ~ ~ 7-,.a17sacr,on on power- ~ ~ ~ ~ ~ ~ l vol. ~ . * 7. no.], pp. 302-309. January 1992.

surge A

152

Authorized licensed use limited to: UNIVERSIDADE ESTADUAL DE CAMPINAS. Downloaded on June 30,2010 at 14:59:33 UTC from IEEE Xplore. Restrictions apply.

(51 Furumasu B.C. and Hasibar R.M.. "Design and Installations of 500 -kV back-to-back shunt capacitor banks". IEEE Trunsuction on Power Delivery. vol. 7. no. 2>pp. 539545. April 1992. [6] Esnieraldo P.C V . . Amon J.F., Salgado F.M.. Carvalho A.C.C. and Morais S.A.. "Circuit - Breaker Requirements for Alternative Configurations of A 500-kV ' ransmission I System". IEEE Irmisactioti on Poiver Delrvety. vol. 14. no. 1 . pp. 169-1 75. January 1999. 171 D. Woodford. In~roduction I O PSCAD V3, Manitoba. Canada: Manitoba HVDC Research Centre Inc.. 2001.
[ 81 A. Greenwood, Electrical 7iansienrs in Power S v s ~ e t n ~ , New York: Wiley. 1991.

155

Authorized licensed use limited to: UNIVERSIDADE ESTADUAL DE CAMPINAS. Downloaded on June 30,2010 at 14:59:33 UTC from IEEE Xplore. Restrictions apply.