ieee schavenmaker 2000a (1)

8/15/2019 Ieee Schavenmaker 2000a (1)

1/5

Digital Testing

of

High4oltage

Circuit

Breakers

circuit hreaker

is a

switching device that the

American National Staiidarcls Institute (ANSI)

defines as:

A

mechanical switching device,

capa

ble

of making, carrying,

ancl

breaking currents under nor-

mal circuit conditions and also making, carrying for a

Pieter J j , Schauemaker,i Lou uan der Suis ;

Rent+

PP Smeets VikrorKert&sz2

ence but also an art. Because of the

complex

phenomelia

involved,

circuit breaker

prototypes

have to

be

verified

by

practical tests in the laboratory.

In

high-pnwer labora-

tories,

the ability of circuit breakers to interrupt shart-

circuit currents

is

verificd in

test

circuits, which

are

in

intended to

operate

que

1 al

t

h

oug

h

infre-

some

types

are

suitable

for fre-

quent operation.

High-voltage circuit breakers play an impartant rolc in

transmission and distribution systems. They must clear

faults

and

isolate

faulted

sections rapidly and

reliably. In

short, they must

possess

th e followiiig qualities:

In

closed position, they

are good

conductors

In

open

position, they are excellent insulators

w They

can

close a

shorted circuit quickly and

safely

without unacceptable contact erosion

They

can interrupt a

rated

short-circuit

current, or

lower

current,

quickly without generating an

abnor-

mal voltage.

The only

physical mechanisiii that

can change

in

a

short

period of time from

a

conducting to an insulating state at

a

certain voltage

i s the arc.

It

is

this principle on w h i ch

all circuit breakers are Iiased.

The first circuit breaker

was

developed

by J.N.

Kel-

man in 1901. It was

the

predecessor

of

the

oil

circuit

breaker and capable of interrupting a short-circuit cur-

rent

of

200 to

300 A in a 40 kV

system.

The

circuit

hreak-

er

was

macle u p

of t w o

wooden barrels containing

a

mixture of nil

and

water, in which the contacts

were

immersed. Since then, circuit breaker

design

has

uncier-

gone

a

reinarkable dcvelopment. Nowadays. one

pole

of

a circuit breaker is capable of interrupting ti3

kA

in a 550

1 V

nctwork, with SF,,gas as the arc quenching medium.

Still,

the design of a circuit Iweakcr is not only

a

sci-

Delft l ln ivcrsity

of

Techriolnfiy

KEMA Iii~h-I'(JwerLaboratory

fact lumped

clement representations

of

the

power sys-

tcm. These test circuits must produce the correct wave-

forms

for

t h c

(short-circuit) current as well

a s

for

t h e

voltage

that

strikes th c circuit breakcr immediately after

the breaker has interrupted the test current. The wave-

forms of current and voltage to which the test object

is

subjected are laid down

i n ANSI and

lnternatioiial Elec-

trotechnical Commission (KC) standards.

These

stan-

dardized waveforms represent 9OX

of the

possible fault

conditions in the real system.


2/5

Circuit reaker Switching and Arc

Modeling

The switching action, the basic function oi thc circuit

breaker, refers

to t h e

change from conductor to insula-

tor at a certain voltage.

Be€ore

nterruption, the (short-

circuit) current flows through the arc channet

of

the

circuit breaker.

Bccause of

the nonzero resistance

of

the

arc channel, this short-circuit current caiises a voltage

across the colitacts

of

the circuit breaker: the arc volt-

age. The arc

behaves

as a nonlinear resistance. Thus,

both arc voltage and arc current cross the zero-value at

the same time instant.

f

the arc is coded sufficiently at

the time the current

goes

through zero, the circuit break-

er interrupts the current,

because

thc electrical power

input

is zero.

During current interruption, the arc resis-

tance increases

from

practically zero to alinnst infinite in

microseconds. Immediately aftcr current interruption,

t h e transient recovery voltage builds u p across the cir-

cuit breaker. As

the

gas

mixture

in

the interelectrodc

space does

not change to

a completely

insulating s tate

instantaneously, the arc resistance is finite

at

that time,

and

a sniall

current

can

flow:

thc post-arc current,

Black-box arc models are mathematical descriptions

of

the electrical properties o l the arc. This type

of

model

does

not simulate th e complicated

physical

processes

inside the circuit breaker liut describes

the

electrical

properties

of

the circuit breaker. Measured voltage

and

current traces are

used to

extract the parameters for the

differential equations describing the nonlinear resistance

of the electrical

arc

for that specific measurement.

Digital Testing

The functionality of high-voltage circuit breakers

is

tcst-

ed in high-power laboratories, Due to the

ncccssary

power

and

the physical size of the equipment, testing is

rather expensive and timc consuming. In order t o obtain

as

much information as possible about the degradation

and

operating limits of t h e circuit breaker from the cost-

intensive tests,

a

project started with the following part-

ners: KEMA High-Power Laboratory,

The

Netherlands;

Delft University of Technology, T h e Nethcrlancls;

Siemcris

RG,

Germany; RWE Encrgie,

Germany;

and

Laborelec

cv,

Helgium. This project

is

sponsored

b y

the

Directoratc GeIieral XII

of

the ELiropcan Commission

in

Standards, Measurements, and Testing Prograiii under

contract

number SMt'4-CT96-2121.

The project is aimcd

at developing digital testing

of

high-voltage circuit break-

ers, i.e.,

a

software product for testing of

I

model of such

a device, once

i t s

characteristic fingerprints are ohtaincd

from rcfinecl ineasurements during standard tests. Digi-

tal testing offers

a

wide range of new possibilities for

users, manufacturers, standardizing bodies,

and

test lab-

oratories for fine tuning circuit hreakcr abilities in rela-

tion with standards and

real

power

systems.

Some

developments

arc:

Evaluation of t he relevance of future standar ds

with respect to real power systems

I Evaluation of the relevance of future standards for

different circuit breaker technologies

and

cxtin-

giiishing media

stimation of the circuit breaker's interrupting limit

Reduction of full-scale testing in high-power labora-

tories

w Iclentificatioii of network topologies that can

pose

spccial difficulties lo a circuit brealrer

Acceleration

of

development of new circuit breaker

clcsigns

Monitoring the aging processes f l t circuit breakers

in

service

Expansion O services for high-powcr laboratories.

The steps followed

so far

to enable digital testing are

described in the following sections. At the end of the arti-

cle, examples

of

digital tcsting are presented.

Measurements and Data

Analysis

I-ligh-resolution ineasurcmcnts

of

current and voltage

in

the

critical period around short-circuit current zero

must

supply the necessary parameters, characterizing

the breakers' behavior.

A

tailor-macle high-frequency

measuring system was realized for this purpose. This

system consists

of

a

number

of battery-pciwcred, single-

chaniicl, 40 MHz, 12 bit AD converters, each storing

the

data temporarily

in

on-board local RAM (2Stik

samples

each). The concept

of

on-site data storage is

necessary

for reaching a maximum overall

sysLern

bandwidth,

Cables

to the current and voltage S ~ I I S O ~ San t h u s be

kept

very

short, and thc sys tem can

operate

011 floating

potential. The arc

voltage i s measured with standard

brnad-band RCR-type voltage dividers; current

is

mea-

surecl with

a

spccial Rogowski coil. After

the remote

RAM

is fillcd, data is transmitted serially through optical

fibers to the proccssirig unit i n the c om m a nd ccntcr. The

greatest challenge with respect t o developing the equip-

rneiit in this application design

lies

in the electromagnet-

ic compatibility, since th e microelectronics

has

to

function in an extremely hostile environment of intense

EM fieldsof various origin.

The system relies heavily

on

digital signal processing

methods for reconstructing the actual voltage

and

cur-

rent signals from the raw scnsor output. On

the

one

hand, t h i s h a s to dc) with the specific frequeiicy

response of the sensors and on the other hand, with cor-

rections needed

fnr

the (reproducible) induced voltages

and capacitive current that distort the measured

signals.

'rests

in various laboratorics have proven that

t h c

sys-

tern can measure post-arc current as small as S

mA,

microseconds after the interruption of many tens

of

kA.

Data analysis software has been produced to carry

out the signal reconstruction practically

o n

line during

the tests (Figure

l ,

and to evaluate

the

performance

O

the tcst object. Even the newest profcssionat multipur-

pose

mathematical or laboratory software

is

not

corn-

petitive to this custom-made software considering

April2000

93


3/5

Short-circuit Current

\

L\

Fails

to Break

I

I 1 m s

. . . . . . . . . . . . . . .

I

.,,

.............

Figure

1.

Vdfage

and

c u r r e n t

mensurrment

data

Z gure

2.

Perameter extraction

software

flexibility and speed in visualizing and data processing

of

practically unlimited amount

of

measured data in

a

user-friendly

way.

After an

extensive

series

ol

the most critical fault

interruption duty

for

circuit breakers (the so-called

“short-line fault,”see the section on “Applications

of

Digi-

tal

Testing”),

a

test database

from

various

types

of

com-

0 50

....

0

10

20 30 40 50

O

70

Sequential

Number

of

Tesl

F u r ~

3.

Degradation oPthe circuit b r e d e r poles

mercially available circuit breakers was

set up.

With this

experimental material, an empirical arc model

Iiased

on

classical arc models

was

validated that

gave very good

coverage

of the

observed processes. From the total num-

ber

a

> E O )

interruption attempts, the result

of

the

atte mpt (failure/success) was predi cted correctly in

mo re

than

90 of

the cases

by

evaluating

the

character-

istics

of

the a rc behavior with the model.

The inodel has a set

of (three) parameters, which are

extracted autninatically during the cvaluation

of

each

test (Figure

2 .

Automated analysis

of

th c collection

of

all

the

parameter

sets (in nther words, the breakers’ “finger-

prints“) obtained from a whole series

of

tests makes it

possible to evaluate various physical quantities as a

function

of

test conditions.

The aim of using this method is t o quantify the hreak-

er performance (the margin M

of

interruption), indicat-

ing how successful

the

breaker passed

the

test (M > 0) or

how

far off it is

from passing

it (M

An

example is given in Figure 3, where the clegracla-

tion

of

the three breaker poles (A, B, and C) is presented

during

a

sequence

of

successive tests.

It

can

be

seen

clearly

that the margin of the breaker decrcascs with

every

test. The rate

of

margin

decay

(among others)

s a

Ineasure of the endurance of the breaker with respect

to

this type o t tests.

0).

Arc Circuit Interaction

Software

At

the final

stage of

the realization of digital testing, mea-

sured arc model parameters will be used as input for the

arc model. Of

course,

this arc model behaves a nonlin-

ear

element

in

the electrical circuit ancl must therefore

be analyzed with

a

dedicated

cnmputer

program. The

analysis

of

arc-circuit interaction iiivolving nonlinear clc-

ments

in

relation to

stiff

differential equations

makes

it

necessary

to perform t h e

calculations with

a

variable

step

size

and

adjustable accuracy

of t he

computed cur-

rents, voltages, and canductaiices.

Because

they

have

fixed step-size solvers, EMTP ancl comparable programs

are less suitable for this purpose

and

thercfore

a n e w

approach, the integration of differential algebraic cqua-

tions (DAE) by means of the backward differentiation for-

mulas @DO

method, has been chnsen

in cleveloping

a


4/5

new softwarc package

lor

electrical transients computa-

tion. This

new

transient program, XTrans, has been

developed at the Delft University

of

Technology especial-

ly lor

arc-circuit interaction studies. The program runs

on

a

PC

with

the

MS-Windows operating system

and

works

fully graphical, as shown

in

Figure

4.

The program

is

in

use

at

several high-power and high-voltage laborato-

ries

in

the

world.

The program makes use

of

libraries that contain

infor-

mation about the behavior of element models. The pro-

gram structure is depicted in Figure 5 .

This

structure has

been realized with object-oriented programming. The

compiled code

of t h e

element

m o d e l s is

placed

i n

dynamic link libraries @Us). he models are, thereforc,

separate from

the

main program, which makes it easy to

create new models and use them in the main program.

A full

working demonstration version of

thc

XTrans

program can

h e downloaded rom the

homepage

of

the

electrical power systems group at the

Delft

University of

Technology,

http://eps.et.tuclelft.nl .

Applications of Digital Testing

Influence of

Parallel

Capacitance

Powerful possibilities with digital testing a re created

when the arc model, validated

as

described

in

the sec-

tion on Measurements and Data Analysis, is coupled

with a circuit analysis packagc. Then, the performanceof

a

circuit breaker, the fingerprints of which were obtained

from real tests, can be estimated in circuits other than

the test circuit.

For examptc, the influenceof various standard substa-

t ion

components on the breakers'

capaliilities

can be

estimated through digital testing.

Here the influence of

a

parallel capacitance

is

calculat-

ed

(for

example,

the

parasitic capacitance

at

a

current

transformer, CT,) in the substation.

In

Table 1, the perfor-

mance of

a short-line fault interruption i s compared in

the presence of two types

O

CTs:

CT 1, having

200

pF of

parasitic capacitancc,

and

C1 2, having 400 pP.

These

CTs can be located near the circuit breaker and remote

(the latter implying a n additional 50

pH

of busbar

between CT and breaker). s a reference, the case with-

out CT has a performance

o f

1 .O.

Table 1

shows

that the difference between the two

types

of

CTs

is

rather sinall when compared to the gain

obtained

by

the CT that

was

installed to the breaker as

closely as possible.

Critical

Line

Length

Determination

One of the most severe currents for a circuit breaker to

interrupt

is the

short-line fault (SLF).

In

the case

of a

short-line

fault,

the short-circuit point

is

on

a

high-volt-

age transmission line

a few

kilometers

away

from

the

breaker terminals. After current interruption, a

very

steep, triangular-shaped

waveform

(with a rate of rise

of

5-10

kV/microsecond) stresses

the

extinguishing

medi-

um between the contacts. The percentage SLF indicates

to what extent the short-circuit current is reduced by the

transmission line, e.g., a short-circuit current

of

40 kA is

reduced to 36 kA in case of a 90% SLF In the

IEC

stan-

dard,

75

and

90% SLF

tests are prescribed.

A s

an example

of

digital testing, the critical line

length,

the

short-line fault percentage that

stresses

the

circuit

breaker

most , will be determined

for a 145

kV,

Figure

4.

X Tr ans

transient program

Libraty

Library I

L

_IL __ I__

I

Figure

5.

Structure o f the

XTrnns

transientprogram

http://eps.et.tuclelft.nl/http://eps.et.tuclelft.nl/


5/5

Artificial

Lino

r - - - - - - - - - - 1

_.

- - -

r

I F 3 -

Model

I'

F-*

Figure 6.

A direct SLF

test

circuit i n the XTrans program

31.5

kA,

SF, circuit breaker.

A

direct SLF test circuit

is

shown in Figure 6, hrce different indicators, active at

different

time

intervals (before current zero, at current

zero, and

after current

zero)

are

used

to quantify

t h e

slress

on

the circuit breaker

model.

Before ciirrent

zero:

the t imc

before

current zero

where

the

arc

resistance

equals

thc surge imped-

ance

of

thc transmission line

t H

=

3

he closer

thc

value

is to current zero,

t he

more severe the

breaker

is

stressed by the test circuit.

At current zero: the

arc

resistance XO. The

lower

the arc resistance

value

at the current zero

cross-

ing,

the

stronger the breaker

is

stressed by the

test

circuit.

After current zero: thc post-arc ener.gy Epa. This

value is t h e integral nf the multiplication of the

small post-arc current

and

the recovery voltage. It

is clear that

only

for successful interruptions an

Epa value

can

be calculated. The higher t h e Epa

value is, the mo r e severe the breaker is stressed by

the

test

circuit.

The actual computation s based on 75 current zero

recordings

of the circuit breaker of whi c h t he

circuit

breaker model parameters

have

been determined. For

cwh sct of parameters, the

stress

at the various short-

h i e

fault percentages is computed. At last, the overall

stress is visualized

which

is shown in Figure

7.

All three indicators

show

that the circuit breaker

model

is

stressed most severely at a

93 SI

,F,whereas a

90 SLF is prescribed in the

IEC

standard. This

shows

that digital testing can

h e

applicd

to use

the information

o1)tainecl

from

laboratory tests

for

the development nf

fuuture

stantlards.

Acknowledgments

The authors are indebted to the Ih-ectornte Gencral X1I

of

the Eurc-

penn

Cominissiori

i n

Stiuidartls, Measurements, aiitl l 'esting Program

for sponsoring. The effort

nf

the other pa rtne rs involvcrl in the

project

is greatly appreciated .

Fo r Further Reuding

K.P.P.

Sinccts, V. Kertesz, Lvaluatinn

of

high-vollage circuit Imakcr

performance with a ncw arc

model," IEE

Pmreed iqs ort Gerierorion

1.5

1,4

5 1,2

g

1 3

1 , l

5 1

, 0,9

.-

5 0,8

a,7

W

0 5

8 86 07

00

09 0 91 92

93

94

95

%

SLF

I _

- RO (R=Z)

Epa

Figure Z riticai h e

ength

determination by m e u m f l fd ig td esthi

Tmrisrriission,orid Uislribuliori,

N.I).H.

Hij l

L van

tlcr

Sluis.

Ncw

approach to th e calculatinn of

electricnl transients

,

Ewopetm Trmmtuhms on E clcctrictd Pou; w h g i -

m e r i n g

(ETEO,

Volume 8 . Niiinlicr

3 .

pp. 1 7 5 -1 7 9 arid

181-IR6.

May/June 1998.

Biographies

Pictcr

H.

Scliavemaker

ohtaiiietl his

MSc

in electrical engineering

from

t lw Delft Univcrsity of

Technulqy

i n

1994.

Alter

gratluntion, he per-

Iorinccl r s nr h OIL puwer system s tate estimation with the Power Sys-

tems 1 .abora to ry . In

1995,

lic s t a r t e d as an app l ica t ion eng ineer

programming Stilistatlon Control S ystems with

ABU

in the Netherlands.

Since 1996, he

hns

bccn

with

thc Power

Systems

I.ahoratnry, where h e

is currently assistant prnfessor.

He

is working 011PhD rcscai-cli

011

cligi-

tal testing

of

high-voltage circuit tireakers within the framework of a

Luropenn project. His rnniii rcscarcli inte rests include power

system

transients antl power system calculations. H e is

a

member

1 KEE.

He

can be read ied by E-mail, P.H.Schavcinakc~i ts . tudelf t ,r~l .

Lou van der S l u i s

ob ta ined h is

MSc

i I i e lec t r ica l eng ineer ing

from th e Delft I lniversity of Technology in 1974. H c julrrcd thc

KEMA

High-Powcr

Labriratory in 1977

as a test

enginccr and

was

irivolvcd

in t h e d e v e lo p m e n t of ;t data acquisition system fur thc Higti-Puwer

Laboratory, cornputer calculations of

test

circuits. ancl the analysis

of

test

da ta by digita1 cornputer. In 1990, l i e became a part- l ime pro-

fessor, a n d , s incc 1992, tic

has beer i

employed a s

n

full-t ime profes -

sor a t t h e llelft

Unlvcrsity

1

T c c t i ~ i u l u g yn

the P n w e r S y s t e m s

Depar tment .

He

is a

senior

m e m b e r

of IKEE arid

cor ivc~ic rof WG

CC-03

of Cigrb

and Circd to s tudy t he t rans ien t recovery voltages

in metliutn and high

vo l tage

networks. IIe c a n h e reached by Bmail,

L.vanderSluisQits.tiidellt.nl.

R e d P.P, Smeetsohtninetl

his

MSc in physic s trorl1 tlie Eiridhrwen

University

of Technology

in

1981. Hc

ohtairwtl the PtiD

degree

in 1987

for research work on vaciiiiin

arcs

at tlic sninc university. From 1983 c

1995, tic was a staIf rriember of the Riiergy Systems Ulvision of thc

Fac-

ulty of Electrical Engineering, Einrlhoven Unlvcrsity. During tlie ycar

1991 hc spcrit

a

sddiatical leave

at

'I'oshilm Corporation's Heavy Appa-

ratus EIigiricerinR Laboratory

i n

Japan.

1111995, 11c

joined KEMA for

R8rD

1 thc High-Power Lilboi-atory.

I

IC is

n

memticr of Cigrd WG 13.04,

the Current Zero Club, Cigr6, and the IEEE

I

le

can

he reached hy 1

niail, r.~~.~~.smeetsQkemn.ril.

Viktor Kcrt6az

otitained his

MSc

in electrlcal engincerlng

irotn

Butlapcst University nf TechndoLy iii 19G6, Phi in 1976, and nSc in

1989. Hc jclirlccl the High-Puwer Laboratory

of t h e

Electrical Power

Research Institute,

Uutlapcst,

as

a rcscarcher arid t a t

erigirieer

i n 1966.

He

ihecainc

a

professor of mathematics at the Budapest Ilniversity o l

'I 'ctchiiology n I979 arid has hcld thc chair of full professor s ince 1989.

He has had

n

scicntiflc contact w i th

KEMR

in the iield

of

circuit breaker

measuring

prolilenis

antl the analysis of arc

pticnnmcna slncc 1978.

He

contributed

as

a m e m l w t n C i gr t W(; 13.01 (arc niotlclirig).

He can

tle

reached by

L m a i l ,

kertes7.vQelender.liu.

ieee schavenmaker 2000a (1)

Documents