IEEETransactionson PowerDelivery Vol. 6, No.

2, Apll 1991 54 1
DEVELOPMENT OF A CURRENT DETECTION TYPE CABLE FAULT LOCATOR

Mitsuharu Komoda Mitsugu Aihara

Takao Kawashima Yasutaka Fujiwara
Masayoshi Arakane Junichi Shinagawa
Chubu Electric Power Co., Inc. Showa Electric Wire & Cable Co., Ltd.
Nagoya, Japan 1-1-2 Odasakae, Kawasaki, Japan

ABSTRACT The present instrument developed by the authors

utilizes a pulse radar method, detecting a traveling
The paper deals with a new type of cable fault wave current and providing a digital display of the
locator. Rapid fault location and repair is required distance to the fault. Since a high voltage pulse is
in case of a cable line breakdown. The Murray applied to cause an electrical discharge at the fault
Bridge, which is the most popular fault-locating point,- the method is applicable even if the
method, is difficult to apply if one of the end resistance at the fault point is high, and can be
terminal of the cable line is connected to a GIS. used for branched or cross-bonded lines.
Moreover, the conventional pulse radar method is This instrument is applicable to cable lines of
difficult to apply if the cable line is branched o r voltage 6.6 kV to 33 kV.
crossbonded. In these voltage classes, the applied pulse voltage
The authors have developed a fault locating does not damage the sound section of the cables.
instrument based upon traveling wave current
detection. Measurement is automated, and the 2. Principle of fault location
distance to the fault point is displayed by merely
pushing a switch. In a branched or crossbonded line, In the initial stage, the faulty phase is
the fault point can be located without interference identified by using conventional methods, measurement
from the branched or the crossbonded point. Tests of insulation resistance o r application of DC
performed in both real and simulated lines confirmed voltage. Then, the present locator can be used.
the practical utility of the present instrument. Figure 1 shows the principle of the present fault
locator. High voltage pulses are simultaneously
Kevwords: Fault location, underground cable, pulse applied to the cables of the sound and faulty phases,
radar then the sum (I1 + 12) and difference (I1 - 12) of
the traveling wave currents entering the two phases
1. INTRODUCTION are detected by high frequency CTs (CT1 and CT2).
Figure 2 (a) shows the situation immediately after
When accidents such as breakdown or burnout occur the application of the pulse. Since the waveforms
in cable lines, the fault must be promptly and and magnitudes of I1 and I2 are identical, a signal
accurately located in order to minimize the time appears in CT1 but not in CT2. The output of CT1
required for restoration of service. Fault location actuates the time counter.
has generally been accomplished by the Murray loop or When the applied pulse reaches the fault, a
pulse radar methods, but in the case of XLPE cables, discharge occurs, and a reflected wave is generated
the resistance at the fault is often very high, which (Figure 2 (b)). As shown in Figure 2 (c), this
makes detection by the above-mentioned methods reflected wave, upon returning to the measurement
difficult. Moreover, concomitantly with the terminal, splits into a component which turns back
introduction of GIS, the end terminals of cable lines into the sound phase, a component which flows into
have been fully enclosed in metal cases, hence the the pulse generator and a component which reenters
terminals available for connection to the cable of the faulty phase.
the leads required for fault location have become CT2 then detects the difference of the reflected’
severely restricted. Consequently, methods such as wave and the output of CT2 stops the time counter.
the high voltage Murray loop procedure, whereby the The time T elapsed between the application of the
three phases are shorted at the remote end terminals high voltage pulse and the appearance of the
and a high voltage imposed, are now difficult to reflected wave at the detection terminal is measured
apply * by a time counter (resolution Ions), and the distance
Many pulse radar type locator^!^"^' are applicable to L to the fault is then calculated from the previously
cable lines of simple configuration. measured propagation velocity v in accordance with
However, no conventional locator can efficiently the formula
locate the fault if the cable line includes the L = v.r/2.
above-mentioned GIS connected terminals, branched In the case of branched lines, reflected waves are
lines o r crossbonded joints. also generated at the branch point, as shown in
The authors have previously developed a digital Figure 3 (b). However, being formed in the same
cable fault locator‘6),but this instrument was liable manner in both the sound and the faulty phases, these
to pick up noises at the voltage dividers. waves arrive at the measuring terminal
simultaneously, as shown in Figure 3 (c). These
waves then split into three components, one flowing
into the pulse generator, one re-entering the other
phase, and the reflected wave. CT2 does not detect
these waves because they are of the same,-magnitude,
and wave form, and identical 1 1 1 Dsth the sound and
faulty phases. The time counter does not stop till
the reflected wave from the fault arrives at the
measuring terminal.
In the case of a cross-bonded line, one more lead
for application of the pulse is connected with the
other sound phase, and the same pulse is applied to

0885-8977Dl~CMm1$01 .OO@ 1991 IEEE

542

all three phases, therefore, the waves reflected from 3 . Specifications of fault locator
the cross-bonded points mutually cancel and the waves
reflected from the fault point alone can be detected The specifications and external appearance of the
without interference. present instrument are shown by Table 1 and the
In many cases, v is 150m/p to 200m/ps. Then, the photograph in Figure 4 , respectively. The fault
time resolution of the time counter corresponds to a location procedure is as follows.
cable length of 0.75111 to 1.0m. Therefore, the
distance resolution of this fault locator is actually 3.1 Calibration of wave propagation velocity
less than Im. However, since zis influenced by the ( 1 ) Connect one lead to each of the sound phases.
wave distortion and the flashover delay at the fault, (2) At the remote end terminal, ground the core of
a difference of Im is not meaningful. Therefore, the one of the phases which have been connected to the
locator actually displays minimum distance increments leads.
of 10m.

Highvoltage generator Pulse generator

(time difference T)

Counter
Propagation velocity
(input)

Fig. 1 Principle of fault locator

O Signal output
(e) Immediately &er applicationofhigh voltage pulse
fa) Immediately aRer applicationofhighvoltage pulse

4
H1:
T : il-12jo
O Signal output
0 Signaloutput
(bl Formationofreflected waveatfaultpoint f b ) Formationofreflectedwave at branched point

T
11-121 0
-
Signal output
(c) Arrival ofwave reflected a t fault point le) Arrival ofwave reflected a t branch point

Fig. 2 Reflection of high voltage pulse and arrival Fig. 3 Reflection of high voltage pulse at branch
of reflected wave at detection terminal point and signal outputs detected in CTI and
CT2.

Table 1 Specifications of fault locator

Charging voltage
I Calibration mode 3kV
Location mode 30kV (Negative Polarity)

ImDressed current waveform I Wave front 1 us Wave tail 5-10 us

Distance display 4 digit(99.99km) Resolutiou O.Olkm

1115mm L X 775mm W X 680mm H; weight 13Okgf

Dimensions and weight

210min L x 350n11n W X 1OOinrnH: weight lOkgf

Fig. 4 Photograph of fault locator
 543

(3) Set the switch of the fault locator to Two types of faults were created, i. e., by short-
CALIBRATION, press the CHARGE button, and after circuiting the core with the shielding and by
charging is completed, press the DISCHARGE button. inserting a discharge gap of variable length.
(4) If the displayed distance differs from the Even discharge gaps 3mm in length (discharge
actual length of the line, rotate the propagation voltage approximately 10kV) were successfully
velocity knob through several turns, then repeat located. In all such cases, the results of
operation (3). automatic location were 30-40111 greater than the
Repeat this procedure until the displayed actual distance.
distance is equal to the length of the line. This error is of the same order whether the
fault is grounded or connected to a discharge gap.
3.2 Fault location Therefore, the cause of this error is not the
(1) Transfer one of leads connected in the statistical time lag of the discharge at the fault.
calibration procedure to the faulty phase. The reflected wave from the fault travels to the
(2) Set the switch of the fault locator to branched point, and then splits into two waves.
LOCATION, press the CHARGE button and after One wave, which enters the short branch, is soon
charging is completed press the DISCHARGE button. reflected at the open end terminal, and this wave
(3) The number which appears on the display is superimposed upon the other wave which travels
represents the distance to the fault. to the measuring terminal. Thus, the wave front is
Thus, operation of the device is extremely blunted and the trigger signal from CT2 is delayed.
simple. Nevertheless, this example demonstrates that the
present instrument permits fault location even in
4. Verification tests branched lines with complicated cable
configurations. However, if the flashover voltage
The present instrument was tested with both actual is higher, a statistical time lag may occur, thus
and simulated lines under various conditions, as causing an error.
indicated in Table 2. Each case will be separately
described below. (d) Case 4
This was a line composed of several types of
(a) Case 1 cable of varying sizes. The grounding points were
This test was performed on a simulated line with simulated by driving in nails up to the cable core
a 6.6kV XLPE cable. At each grounding point, the at various points along the line. The
core and shielding layer of the cable were short- configuration of the line was simple, and accurate
circuited. Accurate location of grounding points location was possible.
over distances ranging from 600m to 1800m was
possible. (e) Case 5
A simulated line was tested in Case 2, but in
(b) Case 2 the present case the tests verified that reflected
Three coaxial cables of characteristic impedance waves from cross-bonding points are not generated
50Q were stranded together and covered with copper even in actual lines. Since the line was in
tape. The copper tape simulated the ground return regular operation, simulated fault points could not
circuit of an actual line. The cable shielding of be created. AC voltages in the tested line were
each phase were cross-bonded every 100m. The cable induced by other nearby circuits. When such
core and shielding were short-circuited at the induced AC voltages were encountered, the present
grounding point. In case of crossbonded lines, apparatus was used with the cores of all three
automatic location of faults has not been possible phases grounded at the remote terminal, whereupon
with the conventional pulse radar method, but circulating current was sometimes found to flow for
faults were precisely located with the present an extremely short time when performing the fault
instrument. location procedure with the loop circuit composed
of the locator and the cable. However,
(c) Case 3 disconnecting at least two phases of the cores at
This was a branched line composed of an SL cable the remote terminal without grounding was found to
combined with XLPE cable of two conductor sizes. suffice as a countermeasure in such instances.

7 Top : CTI output

Bottom : CTZoutput
Faultvohase

GcG_
5W3 Top : currcntwaveformsof
v Sound phase faulty and sound phases
Bottom : differenceoftwo waveforms
A. Case 1: branched line

-
D. Case4eablesofvarioussizes

CTI output
Since no fault points were
Top : CTI output present. no CT2output appeared.
Bottom : CT2 output ZWJd
5dd
E.Case6 cross-bonded line

C. Case 3: branchesanddiasimilar cables

Fig. 5 Examples of waveforms obtained for each case

in verification tests
544

0 0
I

i
000000
?cuc??qu!
000000

000089

* *

N CI *7
> E E E
E E E
000 ~
00o 0m0 0
o o
m m m 44ElEl
r r N d

1
8

4 El
 545

Figure 5 shows representative waveforms obtained in and cross-bonded lines.

each of the cases described above.
REFERENCES
In case of a high impedance fault, if the flashover
voltage of the fault is higher than the arriving 1 ) W.E. Anderson, J.D.Ramboz, A.R. Ondrejka,
pulse voltage, then the fault does not form a short The detection of incipient faults in transmission
circuit. However, when the reflection from the open cables using time domain reflectometry techniques:
circuit end is superimposed, the fault may form a Technical challenges. IEEE Trans. PAS Vol. PAS-
short circuit. In this case, the located distance to 101, N O . 7 , pp.1928-1934, 1982
the fault will be significantly in error. To avoid 2)W.L. Weeks, J.P. Steiner, Instrumentation for the
this error, it is necessary to lower the flashover detection and location of incipient faults in power
voltage by burning the fault with DC current. cables. IEEE Trans. PAS Vol. PAS-101, No. 7,
The authors are now studying improvements of the pp .2328-2335, 1982
locator which will permit automatic calibration of 3)W.Rietz, Fault location in power cables. Wire
the error. Industry Vol. 52, No. 619, pp. 415-416, 418, 1985
4)P.F. GALE, The application of digital electronics
5. CONCLUSION to power cable fault location and testing. Prepr.
Pap. Electr. Energy Conf.Vo1. 1980 pp. 72-79, 1980
A cable fault location instrument based upon 5)P.F. GALE, Practical methods of fault location for
detection of traveling wave currents was developed, high voltage cables. Electr. Eng. Vol. 61, No. 10,
’ different from the conventional pulse radar method pp. 2 1 , 2 4 , 26, 28, 1984
based upon voltage detection. The new locator was 6 ) Y . Suzuki et al., Development of a digital cable
tested on various types of lines. The results fault locating instrument, Proceedings of Jicable
verified that the device operates without 8 7 , €33.3 (1987)
interference even in lines with complicated branches

Mitsuharu Komoda was born in Mitsugu Aihara was born in Tokyo,

Nagoya, Japan, on April 18, 1939. Japan, on September 9 , 1938. He
He received the B.E. degree in received the B.S. and the M.S.
electrical engineering from Nagoya degrees from Tokyo Institute of
University, Japan in 1962. The Technology, Japan, in 1964 and
same year, he joined Chubu Electric 1966, respectively. In 1966, he
Power Co., Inc. He was engaged joined Showa Electric Wire & Cable
mainly in design, construction, as Co., Ltd., Japan, where he has been
well as research and development of engaged in research and development
transmission lines and substation on dielectric phenomena, insulation
until 1988. He joined PSE course design and diagnostic methods of
of GE in 1969-1970. Now he is a high voltage power cables. Now,
Deputy General Manager of the Electric Power Research he is a general manager of electric power research &
& Development Center. development section.
He won the Japan OHM Awards in 1976 and the Prize of He won the Japan OHM Awards in 1982, and the Paper
Progress, I.E.E. of Japan in 1987. Mr. Komoda is a Award from the I.E.E. of Japan in 1989. Mr. Aihara
member of I.E.E. of Japan. is a member of I.E.E. of Japan.

Takao Kawashima was born in Nagoya, Yasutaka Fuiiwara was born in

Japan, on January 6 , 1947. He Okayama, Japan, in 1942. He
received the B.E. degree in received the B.S. degree in
electrical engineering from Nagoya electrical engineering from Tokyo
Institute of Technology, Nagoya, University in 1966. In 1966, he
Japan in 1969. The same year, he joined Showa Electric Wire & Cable
joined Chubu Electric Power Co., Co., Ltd., and had been engaged in
Inc. He was engaged mainly in R & D of superconducting and cryo-
construction of substation and resistive cables for ten years. At
overhead transmission lines until present time he is engaged in the
1988. Then he has been with the development of UHV XLPE cables and
Electric Power Research & Devel- their accessories.
opment Center. He is a senior research engineer of Mr. Fujiwara is a member of the I.E.E. of Japan.
Electric Power Engineering Research Section.
He won the Japan OHM Awards in 1984. Mr. Kawashima Junichi Shinagawa was born in
is a member of 1.E.E of Japan. Miyagi Prefecture, Japan on August
17, 1949. He received the B.S. and
Masavoshi Arakane was born in the M.S. degrees in electrical
Ohita, Japan, on June 2 2 , 1939. He engineering from Tohoku University
graduated from Ohita Technical High in 1973 and 1975, respectively. In
School in 1958. From 1966 to 1989, 1975, he joined Showa Electric Wire
& Cable Co., Ltd., Kawasaki, Japan,
he was with Electric Power Research
& Development Center of Chubu
where he is now engaged in the
Electric Power Company., Inc. He development of UpV cables and
is working on Engineering of Power analysis of the transient over
Transmission Lines and Lightning voltage on cable sheaths.
Surge analysis. Mr. Shinagawa is a member of I.E.E. of Japan.
Mr. Arakane is a member of I.E.E.
of Japan.