New Method for High Impedance Faults Detection Using Total Harmonic Distortion Properties and Time Variations of Current Waveform *

Int
IJ A
ISS
ISS
ww
N
(Ab
met
dist
(TH
oth
Key
1.
Wh
hig
dow
HIF
fau
con
det
bec
the
HIF
to n
det
haz
F
res
res
con
car
and
cau
har
the
non
sam
phe
S
ternational
APE Volume
SN(online) 2
SN(print) 216
ww.ijape.org
IJ APE Volu
New M
Usin
T
bstract) High I
thod for HIFs
tribution netwo
HD) properties
her methods is H
ywords: High I
Introducti
hen an overhe
gh impedance
wn and touch
Fs is its diffi
ults is not en
nventional ove
tection is dif
cause the mai
ermal and me
Fs due to low
network and th
tection is hum
zards as a resu
Fault impedan
sistance and r
sistance is hig
ntact with the
rbide producti
d other reaso
used the curren
rmonics in cur
e HIFs detecti
nlinear loads
me properties
enomenon sho
Several algor
l Journal o
1, Issue 7, O
161-5055
61-6442
ume 1, Issue 7
Metho
ng Tot
Time V
Department of
Impedance Fau
detection in d
orks such as cap
and time variat
HIFs detection f
Impedance Fau
on
ad conductor i
object such a
hes ground, a
icult detection
nough and ge
er current dev
fferent from L
n purpose for
echanical dam
level fault cur
heir equipmen
man lives’ pro
ult of the arcin
nce in HIFs is
resistance betw
ghly nonlinea
e earth during
ion are due to
ons [2]. Non-
nt waveform d
rrent waveform
ion. However
switching and
s as HIFs an
ould be separa
rithms and tec
f Automati
October 2012
7 October 201
od for
tal Har
Variat
Vahid Na
f Electrical & C
ults (HIFs) dete
distribution netw
acitors switchin
tions of current
from random no
ults (HIFs); Tota
in distribution
as branches of
HIF occurs.
n, because th
enerally cann
vices. The mai
Low Impedan
r LIFs detectio
mages to netw
rrents do not c
nt, but the mai
otection and
ng phenomeno
s total of the e
ween earth a
ar due to cha
the arc exten
the heating ge
-linearity of r
distortion. Thi
ms which are
r phenomena
d capacitors s
nd during HIF
ated from it.
chniques hav
on and Pow
2 PP. 165-17
12 PP. 165-17
High
rmoni
tions o
ayebi, Majid Ga
Computer Engin
v.n
ection has alway
works. The pro
ng, linear and no
waveform in fo
onlinear loads s
al Harmonic Di
n network touc
f the tree or b
Basic proper
he current of
ot be detecte
n purpose for
nce Faults (L
on is preventi
work, howeve
cause much da
in purpose for
prevention o
on [1].
earth resistanc
nd conductor
anges in cond
nsion, tubes si
enerated by th
resistance of
is distortion cr
good indicato
such as linea
switching hav
Fs detecting
ve been resear
wer Engine
73
73 www.ijape.o
-- 165 --
Impe
ic Dist
of Cur
andomkar, Moh
neering, Saveh B
[email protected]

ys been difficu
oposed method
onlinear loads s
four stages for H
switching.
istortion (THD)
ches a
breaks
rty of
these
ed by
HIFs
LIFs),
ion of
er for
amage
HIFs
f fire
e, arc
r. Arc
ductor
ilicon
he arc
fault
reates
ors for
ar and
ve the
these
rched
and
magn
low
spec
techn
trans
[2],
algor
In
THD
four
simu
HIF
way.
2. C
In th
detec
and v
2.1.
The
wave
curre
Four
Tran
direc
eering
org ©Science
dance
tortion
rrent W
hammad Javad R
Branch, Islamic
com
ult, because thei
d can easily dis
switching. This
HIFs detection.
); Time Variatio
proposed fo
nitude of phas
order harmo
trum [7-10],
nique [16], k
sform [18-21]
neural netwo
rithm[33-34].
n this paper the
D Properties a
stages is con
ulation results
among other s
.
Canonicals
his paper, fou
ction. These c
value of curre
THD Extra
first canonica
eform. For ev
ent waveform
rier Transform
nsform (DFT)
ctly obtained w
e and Enginee
e Fault
n Prop
Wavef
Ramezani
c Azad Universi
ir signs are not
scern the HIFs
method has bee
One advantage
ons; Distribution
or the detect
se or natural f
onics, sub-ha
high frequen
kalman filter
, expert system
ork [27-30],
e HIFs detectio
and time varia
nducted. Theo
showed that th
same phenom
Used for H
ur canonicals h
canonicals are
ent waveform
action
al for HIFs det
valuating this
m harmonics
m (FFT). FFT i
) and will pr
with the DFT.
ring Publishin
ts Det
perties
form *
ity, Saveh, Iran
t clear. This pa
s from other sa
en used Total H
e of proposed m
n Network.
tion of HIFs
fault current an
armonics and
ncy spectrum
ring techniqu
m [22-26], fau
fuzzy logic
on through the
ations of curr
retical analys
he algorithm w
mena in a sensi
Hifs Detecti
have been ev
based on THD
caused by HIF
tection is THD
s canonical, in
are extracted
is the same as
roduce the sa
It is the only
g Company
ection
s and
*
.
aper presents a
ame phenomen
Harmonic Distor
method compare
s, including
nd voltage [3-
d low freque
m [11-15], fra
ue [17], wav
ult current flic
[31-32], gen
e extraction of
rent waveform
is and MATL
was able to de
tive and selec
on
valuated for H
D, third harmo
F.
D value of curr
n the first sta
d by using F
s Discrete Fou
ame result wh
difference wh
n
new
na in
rtion
ed to
the
- 6],
ency
actal
velet
cker
netic
f the
m in
LAB
etect
tive
HIFs
onic
rent
age,
Fast
urier
hich
hich
Int
is m
x(t)
S
ì
C
ì
W
x
N
P
i
n
H
X
T
har
ext
I
X
W
X
X
ith
I
one
sw
net
oth
2.2
Th
TH
con
TH
con
inc
TH
it i
ternational Jo
IJ APE Volu
much faster. S
) are calculate
ì
(P) = x
N
n=1
ì
(P) = x(
N
n=1
Where
x(n) is magnit
N is data wind
P is harmonic
iis number of
n is nth sampl
Harmonics am
X
ì
(P) =
_
(S
Then in the s
rmonics extra
tracted.
IEÐ =
X
ì
(E)
X
ì
(F)
X
ì
(E) = _
n
P=
Where
X
i
(P) is RMS
X
i
(F) is RMS
data window
If the current
e of disturban
itching, nonlin
twork. Next c
her same phen
2. Stability o
e second cano
HD. For evalu
nsecutive sam
HD is greater
nsecutive sam
crease of curr
HD is permane
is not perman
ournal of Aut
ume 1, Issue 7
Sine and cosin
ed with (1) and
(n)sin(
(n)cos (
2nnp
N
tude of nth sam
dow length.
c number.
data window.
le of data wind
mplitudes com
ì
(P))
2
+(C
ì
(
second stage,
acted of the
)
)

(X
ì
(P))
2
n
=2

value of Pth
value of the
.
THD is greate
nces such as
near load swit
canonicals are
nomena.
of THD
onical for HIF
uating this can
mple data wind
r than the c
mple data win
rent THD is
ent for nonline
nent for linear
tomation and
7 October 201
ne coefficients
d (2) relations

p
)
mple.
.
dow.
mputed accordi
(P))
2

using the rel
e first stage,


harmonic of
fundamental
er than the con
linear load sw
tching or HIF
e used for sep
Fs detection is
nonical, the c
dows is calcul
considered thr
ndows, it can
permanent. In
ear loads switc
r loads switch
d Power Engi
12 PP. 165-17
s of FFT for s
s.


ing to relation

ations (4), (5
current TH


ith data windo
components o
nsidered thres
witching, capa
can be occurr
paration HIFs
stability of cu
urrent THD f
lated. If the cu
reshold in th
be concluded
ncrease of cu
ching and HIF
hing and capac
ineering
73 www.ijape.o
- 166 -
signal
(1)
(2)
n (3).
(S)
) and
HD is
(4)
(S)
ow.
of the
shold,
acitor
red in
s and
urrent
for M
urrent
he M
d that
urrent
Fs, but
citors
switc
Cu
nonl
and 4
Th
conc
netw

























org ©Science
ching.
urrent THD fo
inear load sw
4 respectively
herefore if inc
cluded that no
work.
Figure 1
Figure
Figure 3

e and Enginee
or linear load s
witching and H
y.
rease of curre
nlinear load s
1. Current THD
e 2. Current TH
. Current THD
Figure 4. Cur

ring Publishin
switching, cap
HIF is shown i
ent THD is per
switching or H
D of linear load
HD of capacitor
of nonlinear lo
rrent THD of H
IJA
g Company
pacitor switch
in Figures 1, 2
rmanent, it can
HIF is occurred
switching
switching
oad switching
IF
APE
ing,
2, 3,
n be
d in
Int
I
sep
this
non
nex
non
2.3
In a
of
but
pro
I
and
Ac
is g
loa

u


























2.4
A
beg
rea
all
Alw
ternational Jo
IJ APE Volu
In the many r
paration HIFs
s canonical is
nlinear loads s
xt two cano
nlinear loads s
3. Third Har
all HIFs, alwa
half of curren
t usually that
operties are sh
In this paper,
d nonlinear loa
ccording to rel
greater than 5
ad switching is
u =
h
3
THD
. 1û
4．Time Var
Always HIF c
ginning of fau
aches its final
HIFs and o
ways current
Figure 5. Ha
Fi
ournal of Aut
ume 1, Issue 7
researches, ra
and nonlinea
s not good for
switching in d
nicals are u
switching sim
rmonic of T
ays current thir
nt THD towar
t is not for n
hown in Figure
these propert
ads switching
ation (6), if α
0% HIF can b
s occurred.
ûû
riations
current is abou
ult occurrence
value [37]. Th
obtained from
variations to
armonic spectru
igure 6. Harmo
tomation and
7 October 201
andomness can
ar loads switc
r HIFs detecti
distribution fee
used for sepa
multaneously.
THD
rd harmonic is
rd other harm
nonlinear load
es 5 and 6.
ies are used f
and therefore
(ratio of third
be occurred; o

ut 60 per cent
and gradually
his property is
m tests in dif
ward time ar
um of nonlinear
onic spectrum o
d Power Engi
12 PP. 165-17
nonical is use
ching [11, 15]
ion due to ran
eders. In this p
aration HIFs
s formed more
monics [26, 35
d switching. T
for separation
e HIFs are dete
harmonic to T
otherwise nonl

of its final val
y over 3 to 4 c
s common bet
fferent condit
re increased i
r load switching
of HIF
ineering
73 www.ijape.o
- 167 -
ed for
], But
ndom
paper,
and
e than
, 36],
These
HIFs
ected.
THD)
linear
(ó)
lue in
cycles
tween
tions.
in the
first
varia
nonl
Figu
In
and n
β>0
cycle
occu
[ =

























g

org ©Science
few cycles aft
ations are zer
inear loads s
ures 7, 8 and 9
n this paper, th
nonlinear load
or in other w
es after the
urred in netwo
=
JIrms
Jt

Figure 7.C
Figure 8. Cur

e and Enginee
fter the occurre
ro or variable
switching. Th
9.
hese propertie
ds switching.
words βis stric
occurrence
ork.

Current variation
rrent variations
nonlinear lo

ring Publishin
ence of HIFs,
e (Increase a
hese propertie
es are used for
So if accordin
ctly increasing
of disturbanc

ns at relaying p
at relaying poi
oad switching
IJA
g Company
but usually th
and decrease)
es are shown
r separation H
ng to relation
g in the first
ce, HIF can
(
oint for HIF
int for constant
APE
hese
for
n in
HIFs
(7),
few
be
7)
Int























R
inc
3 o
sim
3.
Im
Th
phe
cap
pro
In
cal
har
cur
TH
wil
beg
gre
oth
In
che
win
nex
wh
det
alg
sam
ternational Jo
IJ APE Volu
Reason of si
creasing reliab
or 4 for the phe
milar to HIFs,
Proposed
mplementat
is algorithm is
enomena such
pacitors switc
ocedure of the
this algorithm
lculated. In t
rmonics is ex
rrent THD is
HD is greater
ll go to next
ginning of the
eater than the
herwise algorit
the next stag
ecks stages l
ndows and if t
xt stage. In the
hich saved in th
tected. Otherw
gorithm. “i” is
mple data win
Figure 9.Curre
ournal of Aut
ume 1, Issue 7
imultaneous u
bility of propo
enomenon exc
HIFs can be d
Det
ton
s to discrimina
h as linear and
ching. Figure
e proposed me
m, after sampl
the next stag
xtracted. Usin
calculated. In
than the cons
stage and oth
algorithm. In
%50, the algo
thm returns to
e, I
rms
saved i
isted for oth
the terms are r
e next stage, β
he memory an
wise, algorithm
a counter that
ndows.
ent variations at
load
tomation and
7 October 201
use of canon
sed method. I
cept the mentio
detected by ot
tection
ate between H
d nonlinear lo
e 10 shows t
ethod.
ling of curren
ge using the
ng the extract
n the next st
sidered thresh
herwise algori
n next stage, αi
orithm will go
o the beginnin
in the memor
er M consec
realized, the a
β is calculated
nd if β is greate
m returns to th
t signifies num
t relaying point
d switching
d Power Engi
12 PP. 165-17
nicals 3 and
f one of canon
oned phenome
ther canonical
Algori
HIFs and other
oads switching
the HIF dete
nt waveform, I
FFT, the cu
ted harmonics
age, if the cu
hold, the algo
ithm returns t
is calculated, i
o to next stag
ng of the algor
ry. Then algo
utive sample
algorithm will
for first four c
er than zero, H
he beginning o
mber of consec
for random non
ineering
73 www.ijape.o
- 168 -
4 is
nicals
ena is
.
ithm
same
g and
ection
I
rms
is
urrent
s, the
urrent
rithm
to the
if α is
e and
rithm.
rithm
data
go to
cycles
HIF is
of the
cutive
Th
with
vary
the p
solvi
Figu
have
Δh
3i
=
exit
using
nlinear

org ©Science
his algorithm
hout harmonic
y with the entry
proposed algo
ing this probl
ure 11. As can
e been used
=|h
3i
-h
3ref
|). TH
any non-line
g the modified
Fig

e and Enginee
is able to HI
c variations. B
y and exit of n
orithm will no
lem, the prop
n be seen, inst
their variat
HD
ref
and h
3ref

ar load in ne
d algorithm in
gure 10.HIF det

ring Publishin
IFs detection
But harmonics
nonlinear load
ot able to HIF
posed algorith
tead of using
tions (ΔTHD
f
will be updat
etwork. HIFs
n all condition
tection procedu
IJA
g Company
only in netwo
s of network w
ds in network
Fs detection.
hm is changed
the THD
i
and
D
i
=|THD
i
-THD
ted with entry
will be detec
ns.
ure
APE
orks
will
and
For
d to
d h
3i
D
ref
|,
and
cted

Int

ternational Jo
IJ APE Volu
ournal of Aut
ume 1, Issue 7
tomation and
7 October 201
F
d Power Engi
12 PP. 165-17
Figure 11. HIF
ineering
73 www.ijape.o
-- 169 --
detection gene

org ©Science
ralized procedu

e and Enginee
ure

ring Publishin
IJA
g Company
APE

Int
4.
4.1
Mo
Em
Fig
wh
and
rep
asy
vol
cur
wh
For
cur









A
dif











In
and
tha
HIF
ternational Jo
IJ APE Volu
Modeling
1. HIF Mode
odel of HIF m
manuel model
gure 12. The
hich represent
d the distribu
present the f
ymmetric faul
ltage is greate
rrent flows to
hen the line vo
r values of th
rrent flows [10
According to F
fferent intensit
n this paper, fo
d better simul
at is shown in
F phenomeno
Figu
F
ournal of Aut
ume 1, Issue 7
el
makes up an im
of HIF used
model includ
the inception
ution line. T
fault resistanc
lt currents to
er than the po
oward the gro
oltage is less t
he phase volta
0].
Figure 13, whe
ties and simul
or consideratio
lation, Emanu
Figure 14. A
n is shown in
ure 12. Emanu
Figure 13. Arc
tomation and
7 October 201
mportant part o
d in this pape
des two DC s
arc voltage be
The two resis
ce: unequal v
be simulated
ositive DC vo
ound. The fau
than the negat
age between
en a HIF occu
ltaneously be c
on of this prop
uel model with
typical curren
Figure 15.

uel model of HIF
s generated by
d Power Engi
12 PP. 165-17
of fault simula
er is shown i
ources, V
p
an
etween trees o
stances R
p
and
values allow
d. When the p
oltage V
p
, the
ult current rev
tive DC voltag
V
p
and V
n
no
rs, several arc
created.
perty in HIF m
h multi-arc is
nt waveform u
F
HIF
ineering
73 www.ijape.o
-- 170 --
ation.
n the
nd V
n
,
or soil
d R
n
,
s for
phase
fault
verses
ge V
n
.
fault
c with
model
used,
under

















4.2.
4.3.
Nonl
1.
2.
Lo
4.4.
The
show
the
param
are s









org ©Science

Nonlinear L
linear loads ar
One phase re
Three phase d
oads are used
Distribution
power system
wn in Figure 1
test network
meters and lo
shown in Table
Figure 14
Fig
Figure 16.

e and Enginee
Load Model
re used in this
ectifier
diode pole
in constant an
n Network M
m studied is i
6. Using of M
is simulated
oads character
es 1 and 2, res
4. Emanuel mod
gure 15. Curren
One-line diagr

ring Publishin
l
s paper.
nd random mo
Model
in voltage lev
MATLAB powe
d. Circuit tra
ristics of dist
spectively.
del of HIF with
nt waveform of
ram of distributi
IJA
g Company
odes.
vel 20 kV wh
er system tool
ansmission li
tribution netw
h three arcs
f HIF
ion network
APE
hich
box,
ines
work

International Journal of Automation and Power Engineering IJAPE
IJ APE Volume 1, Issue 7 October 2012 PP. 165-173 www.ijape.org ©Science and Engineering Publishing Company
- 171 -
Table 1. Circuit Line Parameters
Positive sequence resistance R
1
, Ω/km 0.01273
Zero sequence resistance R
0
, Ω/km 0.3864
Positive sequence inductance L
1
, mH/km 0.9337
Zero sequence inductance L
0
, mH/km 4.1264
Positive sequence capacitance C
1
, μF/km 0.0120
Zero sequence capacitance C
0
, μF /km 0.0075
Table 2. Loads Characteristics
Load number S ( MVA ) Cos ( phi )
1 1.75 0.85
2 1.4 0.85
3 2 0.9
4 0.87 0.86
5 0.375 0.8
5. Results
100 cases of HIF, 80 cases of nonlinear load switching, 80
cases of linear load switching, and 80 cases of capacitor
switching on the distribution network are tested. Algorithm
has been tested for M = 6 on the mentioned phenomena, which
HIFs are well separated from other same phenomena and are
detected. Some of the results are shown in the Table. 3. These
are shown that increase of current THD is permanent and
always current THD is greater than considered threshold for
nonlinear loads switching and HIFs, but that is not for linear
loads switching and capacitors switching. α and β are used for
distinction between HIFs and nonlinear loads switching. α
value is greater than 50% for HIFs and is smaller than 50% for
nonlinear loads switching. β value is greater than zero for HIFs,
but its value is zero or variable( positive and negative) for
nonlinear loads switching.
6. Conclusion
The proposed method provides a new method for HIFs
detection using FFT. The current harmonics are extracted by
using FFT and calculated current THD. Then HIFs are
detected by using four canonicals of current waveform: 1.
THD extraction, 2. Stability of THD, 3. Third harmonic of
THD, 4. Time variations. The proposed method has been
tested under many variations of networks operating conditions
including capacitors switching, linear and nonlinear loads
switching which are similar to HIFs. The results obtained from
the proposed method showed that HIFs can be detected and
easily discriminated from other same phenomena.
Table 3.Results of Proposed Algorithm for N = 1, M = 6, Threshold = 0.05
Phenomenon i = 1 i = 2 i = 3 i = 4 i = 5 i = 6
Linear load switching ΔTHD (%) 2.4 0.02 0.01 0.01 0.01 0
Capacitor switching ΔTHD(%) 3.44 0.24 0.09 0.06 0.04 0.02

Constant nonlinear load Switching
ΔTHD(%)
α(%)
I
rms
(Relaying point) (A)
3.52
16.47
351.2
3.38
14.49
351.2
3.37
14.83
351.2
3.37
15.13
351.2
3.37
15.12
351.2
3.37
15.13
351.2

Random nonlinear load Switching
ΔTHD(%)
α(%)
I
rms
(Relaying point) (A)
3.26
11.35
339.46
3.32
10.8
351.35
3.19
11.69
343.81
3.46
12.35
363.37
3.24
11.26
348.51
3.01
11.14
331.14

HIF 1
ΔTHD(%)
α(%)
I
rms
(Relaying point) (A)
3.38
97.34
344.37
3.76
97.34
348.19
3.94
97.46
355.97
4.44
97.75
359.5
4.65
97.6
361.4
4.51
97.43
359.9

HIF 2
ΔTHD(%)
α(%)
I
rms
(Relaying point) (A)
4.16
89.4
361.3
4.31
90.63
368.8
4.56
90.71
374.53
4.96
92.2
382.16
4.85
91.9
381.85
4.81
92.25
379.66

HIF 3
ΔTHD (%)
α(%)
I
rms
(Relaying point) (A)
6.14
94.6
370.21
6.69
94.81
381.62
6.87
94.78
393.57
7.08
95.15
408.46
7.12
94.35
410.47
7.45
96.68
414.53
International Journal of Automation and Power Engineering IJAPE
IJ APE Volume 1, Issue 7 October 2012 PP. 165-173 www.ijape.org ©Science and Engineering Publishing Company
-- 172 --
References
[1] IEEE Power System Relaying Committee Working Group D15
Report, “High impedance fault detection technology” March 1,
1996. Available:
http://grouper.ieee.org/groups/td/dist/documents/highz.pdf.
[2] A.F.Sultan, G.W.Swiff, and D.J.Fedirchuk,”Detecting arcing
downed-wires usingfault current flicker and half-cycle,” IEEE
T-PWRD, Vol. 9,No.1,Jan 1994, pp.461-470.
[3] AL-Dabbagh, M.Technisearch Ltd. –RMT University, “High
impedance fault detector”, Australian Patent PL3451, July 1992.
[4] A. Nikander, P.Jarventausta,”Methods for earth fault
identification and distance estimation in a compensated medium
voltage distribution network”, Proceedings of the energy
management and power delivery, pp.595-600, 1998.
[5] F. Ruz, A. Quijano, E. Gdmez,”DSTRP: A NEW Algorithm for
High Impedance Fault Detection in Compensated Neutral
Grounded M. V. Power Systems,” ETEP Vol.13, No. 1,
Jauary/Februrary, pp. 23-28, 2003.
[6] Hou. Daqing, N. Fischer,” Deterministic High-Impedance Fault
Detection and Phase Selection on Ungrounded Distribution
Systems”, IEEE/IAS Industrial & Commercial Power Systems
Technical Conference, ICPS 2007. 6-11 May 2007, pp. 1-10.
[7] D. I. Jeerings, I. R. Linders,”A Practical Protective Relay For
Down Conductor Faults”, IEEE Transaction on Power Delivery,
Vol. PWRD-6, No. 2, pp.565-574, April 1991.
[8] J. Reason, “Relay Detects Down Wires By Fault Current
Harmonics”, Electrical World, Vol. 208, No. 12 December 1994
pp 58-59.
[9] R. Das, D. Bayoumi, System For Detection of High Impedance
Fault,” 19
th
International Conference on Electricity Distribution
Vienna, 21-24 May 2007, Paper – 0876, CIRED 2007 Session 3,
Paper No 0876, pp.1-4.
[10] T. Cui, X. Dong, Z. Bo and S. Richards,” Integrated Scheme for
High Impedance Fault Detection in MV Distribution System”
IEEE/PES Transmission and Distribution Conference and
Exposition, Latin America, 13-15 Aug. 2008, pp. 1-6.
[11] B. M, Aucoin, B. D. Russell,“Distribution high impedance fault
detection utilizing high impedance frequency current
components,” IEEE Transaction on Power Apparatus and
System, Vol. PAS-101, No. 6, June 1982, pp. 1595-1606.
[12] H. Y. Chu, M. T. Chen, C. L. Huang, S. L. Chen, S.S. Yen,”High
impedance fault tests on the Taipower primary distribution
system”, Electric Power System Research, Vol. 19, 1990, pp.
105-114.
[13] M. El-Hami, “A distribution system fault location technique
high frequency spectrum of fault current,” Galway, Ireland,
1994.
[14] B. Don Russell and B. Michael Aucoin, 1996,”Energy Analysis
Fault Detection System”, U. S. Patent No. 5, 512, 832.
[15] R. D. Christie, H. Zadegol, M. M. Habib,” High Impedance
Fault detection in low voltage networks”, IEEE Transaction on
Power Delivery, Vol. PWRD-8, No. 4, October 1993, pp.
1829-1834.
[16] A. V. Mamishev, B. D. Russell, and Carl L. Benner,”Analysis of
high impedance faults using fractal techniques,” IEEE Trans.
Power Systems, vol. 11, pp. 435-440, Feb. 1996.
[17] Adly. A. Girgis, Wenbin. Chang, Elham. B. Makram,” Analysis
of high-impedance fualt generated signals using a kalman
filtering approach,” IEEE Trans. Power Delivery, Vol. 5, No. 4,
Nov1990, pp.1714-1724.
[18] D.C. Chan, T. Wai and X. Yibin, “A novel technique for high
impedance fault identification,” IEEE Transactions on Power
Delivery, vol. 13, No. 3, pp. 738-744, 1998.
[19] L. L. Lai, E. Styvaktakis, A. G. Sichanie,” application of discrete
wavelet transform to high impedance fault identification”, IEEE
Transactions on Power Delivery, Vol.2, pp. 689-693, 1998.
[20] T.M. Lai, L.a. Snider, E. Lo,” Wavelet transform based relay
algorithm for the detection of stochastic high impedance faults,”
Electric Power Systems Research 76 (2006) 626-633.
[21] N. I. Elkalashy, M. Lehtonen, H. A. Darwish, A-M. I. Taaleb
and M. A. Izzularab,” Feature Extraction of High impedance
arcing Faults in Compensated MV Networks.Part I:
DWT-Based Analysis of Phase Quantities”, IEEE Power
Engineering Society Conference and Exposition in Africa,
Power Africa ’07,16-20 July 2007, pp. 1-6.
[22] B. D. Russell, R.P. Chinchali, “ A digital signal processing
algorithm for detecting arcing faults on power distribution
feeders”, IEEE PES Winter Meeting, New York, NY, 1988,
Paper No. 88, WM 123-2.
[23] C.j. Kim, B.D. Russell,” Classification of faults and switching
events by inductive reasoning and expert system methodology”,
IEEE PES Winter Meeting, New York, January 1989, Paper No.
89WM058-9-PWRD.
[24] C. L. Benner and B. D. Riussell, “Practical high impedance fault
detection for distribution feeders,” in Proceedings 1996 Annual
Conference of Rural Electric Power, pp.B2-IVB2-6.
[25] C. L. Benner, B. Russell,” Practical High Impedance Fault
Detection on Distribution Feeders,” IEEE Transactions on
industry Applications. Vol. 33, No. 3, May/Jun. 1997,
pp.635-640.
[26] Y. Sheng and S. M. Rovnyak, “Decision tree based
methodology for high impedance fault detection,” IEEE
Transactions on Power Delivery, vol 19,No. 2, pp.533-536,
2004.
[27] S. Ebron, “ A neural processing strategy for the detection of high
impedance faults”, Master’s Thesis, Electrical and Computer
Engineering Department, NCSU, 1988.
[28] S. Ebron, D. L. Lubkeman, M. White,” A neural network
approach to the detection of incipient fault on power distribution
feeders,” IEEE Transaction on Power Delivery, Vol. PWRD-5,
No. 2, April 1990, pp. 905-912.
[29] Peter B. Snow, Snow, Alexander P. Apostolov and Jefferson D.
Bronfeld, 1996, “Method and Apparatus For Detecting
International Journal of Automation and Power Engineering IJAPE
IJ APE Volume 1, Issue 7 October 2012 PP. 165-173 www.ijape.org ©Science and Engineering Publishing Company
- 173 -
High-Impedance Faults In Electrical Power Systems”, U.S.
Pantent No. 5, 537.
[30] R. Keyhani, M. Deriche, and E. Palmer, “A high impedance
fault detector using a neural network and subband
decomposition,” in Proc. 6
th
International Symposium on Signal
Processing and its Applications (ISSPA), Kuala Lumpur,
Malaysia, 13-16 August, 2001. Vol. 2, pp.458-461.
[31] Jota, F.G., and Jota, P.R.S.: “High impedance fault identification
using a fuzzy reasoning system”, IEE Proc. Gener. Transm.
Distrib., 1998,45, (6), pp. 656-661.
[32] Ruz, F., and Fuentes, J.A. “Fuzzy decision making applied to
high impedance fault detection in compensated neural grounded
MV distribution systems”. Proc. of 7
th
IEE Int. Conf. on
Developments in Power System Protection, 2001,pp. 307-310.
[33] R. Sedighi, M.R. Haghifam, O.P. Malik,” Soft computing
applications in high impedance fault detection in distribution
systems,” Electric Power Systems Research, Volume 76, Issues
1-3, September 2005, Pages 136-144.
[34] N. Zamanan, J. Sykulski, A. K. Al-Othman,” Arcing High
Impedance Fault Detection Using Real Coded Genetic
Algorithm”, Proceeding of the Third IASTED Asian Conference,
Power and Energy Systems, 2-4 April, Phunket, Thailand.
[35] S. R. Nam, J. K. Park, Y. C. Kang, T. H. Kim,” A Modeling
Method of a High Impedance Fault in a Distribution System
Using Two Series Time-Varying Resistances in EMTP”, IEEE,
Vol. 2, pp.1175-1180,2001.
[36] NaserZamanan, Jan K. Sykulski,” Modeling arcing high
impedances faults in relation to the physical processes in the
electric arc” WSEAS TRANSACTIONS ON POWER
SYSTEMS, Issue 8, Volume 1, August 2006.
[37] D. I. Jeerings, I. R. Linders,” Ground resistance-revisited”,
IEEE Transaction on Power Delivery, Vol. PWRD-4, No. 1,
1989, pp. 949-956.









Author Introduction
Vahid Nayebi was born in Iran, on Aug 02, 1987.
He received the B.Sc. degree in electrical power
engineering from the Azad University of
Technology, Saveh, Iran, in 2009. He has been
pursuing M.Sc. degree in electrical power
engineering at Azad University of Technology,
Saveh, Iran. His research interests are Microgrid and power system
protection, power system analysis.E-mail: [email protected].
Majid Gandomkar was born in Saveh, Iran in
1973. He received his B.Sc and M.Sc. degrees in
electrical power engineering from Sharif University
of Technology 1995 and 1997 respectively and PhD
degree in electrical power engineering from
Science and Research Campus, Tehran, Iran.
Currently he is an Assistant Professor with Saveh Azad University of
Technology. His research interests are distribution system, Microgrid
and power system protection.
Mohammad Javad Ramezani was born in Iran,
on July 13, 1985. He received the B.Sc. degree in
electrical power engineering from the Azad
University of Technology, Saveh, Iran, in 2008. He
has been pursuing M.Sc. degree in electrical power
engineering at Azad University of Technology,
Saveh, Iran. He was a power engineer in powerhouse, Tehran, Iran,
from 2008 to 2010. Also, He is a manager of Electrical and
Instrument Group in Iranian Offshore Engineering & Construction
Company (I.O.E.C), from 2010. His main research areas are power
systems fault diagnosis, Microgrid and power system protection and
automatic event analysis.