TELKOM
NIKA
, Vol.12, No
.4, Dece
mbe
r
2014, pp. 75
3~7
6
2
ISSN: 1693-6
930,
accredited
A
by DIKTI, De
cree No: 58/DIK
T
I/Kep/2013
DOI
:
10.12928/TELKOMNIKA.v12i4.300
753
Re
cei
v
ed Au
gust 27, 20
14
; Revi
sed O
c
t
ober 1
2
, 201
4; Acce
pted
Octob
e
r 25, 2
014
Spectrum Comparative Study of Commutation Failure
and Short-Circ
u
it Fault in UHVDC Transmission System
Chen Shi-lo
ng, Rong Ju
n-xiang, Bi
Gui-hon
g, Li Xing-
w
a
ng,
Cao Rui-rui
Schoo
l of Elect
r
ic Po
w
e
r En
gi
neer
ing, Ku
nmi
ng Un
iver
sit
y
o
f
Science an
d
T
e
chnolog
y, K
unmi
ng 65
05
0
0
,
Yunn
an Prov
in
ce, Chin
a, T
e
l.138
88
639
30
5
E-mail: chens
hilo
ng
3@1
26.c
o
m
A
b
st
r
a
ct
When co
mmut
ation fa
ilur
e
oc
curs in UH
VD
C trans
missi
on
system, the transi
ent proc
es
s of DC
voltag
e
and
cu
rrent ar
e s
i
mil
a
r to gr
ou
ndi
ng
short-circuit
fault. In
order to
differentia
te them effectively,
the
pap
er introd
uc
es math
e
m
atic
al morp
hol
ogy
meth
ods to
an
alysis the sp
ec
trum of transi
e
nt current. Base on
Yunn
an-Gu
ang
z
h
o
u
80
0
kV UHV
DC trans
m
i
ss
ion system
, the paper
sim
u
lat
e
s the commutation failur
e
and
DC li
ne s
h
ort-circuit fault
und
er differe
nt fault co
n
d
itio
ns
in PSCAD/EM
T
DC. By mo
di
fied
mor
pho
lo
g
y
filter, the transi
ent sign
al of D
C
(
d
I
) is deco
m
p
o
sed i
n
to six s
c
ales, an
d
mor
pho
log
i
cal c
har
acteristics of
aeri
a
l mode
co
mp
on
ent of
d
I
is an
aly
z
e
d
und
er differ
ent sc
a
l
es. T
h
e
si
mu
l
a
tion
resu
lts s
how
that w
h
e
n
DC l
i
ne
sh
ort-circuit fa
ults occ
u
rs, w
herever
i
n
the
rect
ifier
s
i
de, i
n
th
e D
C
t
r
ans
missi
on
li
n
e
mid
p
o
i
nt or
i
n
the inverter si
de, the
aeri
a
l
mo
de co
mpon
ent of
d
I
have mor
e
hi
gh fre
que
ncy w
e
ight
in
1
d
~
5
d
and
decays
grad
ua
lly; When co
mmutati
on fa
ilur
e
s, w
h
ich are
cause
d
by the
inverter si
de
AC system si
n
g
le-
phas
e gro
und
i
ng fault, ph
ase
to phase fa
ult
,
three phas
e
grou
ndi
ng fau
l
t or the inverter
side transfor
m
er
ratio incr
ease
d
,
the aerial
mo
de co
mp
on
ent of
d
I
have less frequ
ency w
e
ig
h
t
in
6
d
.
Ke
y
w
ords
: UHVDC, commutation failur
e, DC
line sh
ort circu
i
t fault, PSCAD/EMT
DC, morp
hol
ogy
1. Introduc
tion
Comm
utation
failure is on
e of the highes
t pro
babl
e
failures in
UHV
DC tran
smissio
n
sy
st
em
[1].When
comm
uta
t
ion failure
occurs, it w
ill
ca
use
DC voltage sharply de
clinin
g and
DC
curre
n
t sud
d
enly increa
si
ng whi
c
h
will serio
u
sly
affect the po
we
r quality of DC tra
n
smission
system
[2] an
d may
ca
use
the in
co
rre
ct
operation
of
relay p
r
ote
c
tio
n
eq
uipme
n
ts in A
C
syste
m
[3],[4]. Grounding
short-circuit faul
t in
DC transmi
s
si
on sy
stem has si
milar
DC
voltage and
DC
curre
n
t tra
n
sient characte
ristics comp
are
with
co
mmutation f
a
ilure.
The
r
e
f
ore, it i
s
g
r
eat
signifi
can
c
e t
o
the safe an
d stable o
peration
of UHV
DC transmission system [5
] to distinguish
comm
utation
failure an
d DC line short
-
ci
rcuit fault is o
f
great import
ance.
At prese
n
t, the traditio
nal
discrimin
a
tin
g
method
s o
f
commutatio
n
failure in
cl
ude the
minimum vo
ltage dro
p
method, the
minimum
arc
angle
criterio
n meth
od, etc
[6]-[
8
]
.
Non
e
thele
s
s, these meth
o
d
s ta
ke
many
factors
i
n
to accou
n
t
and
can’t disting
u
i
s
h com
m
utati
o
n
failure from DC line sh
ort circuit
[9]. By u
s
ing MATLA
B
/Simulink, referen
c
e [10]
emulates 50
0kV
HVDC tran
smissi
on
syste
m
and de
co
mposes th
e DC
cu
rre
nt into six scale
s
by wavelet m
u
lti-
scale
analy
s
i
s
. Th
e exp
e
riment results point
that t
he
DC
curre
n
t ha
s m
o
re
high
fre
que
ncy
division
s in DC line short ci
rcuit fault and
more lo
w fre
quen
cy divisi
ons in
comm
utation failure
.
In re
cent ye
ars,
mathem
atical m
o
rp
h
o
logy ha
s
be
en wi
dely ap
plied in
po
wer q
uality
detectio
n
, sin
gular poi
nt d
e
tection,
harmonic an
al
ysis of tran
sient
sign
al, fault
diagn
osti
c, re
lay
prote
c
tion a
n
d
fault locati
on, etc
[11].
Whe
n
processing m
u
lti-scale
of sig
nal
s, mathem
atical
morp
holo
g
y
decompo
se
s sig
nal
s in
time do
main,
whi
c
h
ca
n m
a
ke
up
for d
e
ficien
cie
s
th
at
Fouri
e
r tran
sform
can’t full
y describ
e time vari
a
n
t n
on statio
nary
sign
al [12],[13] and
avoi
d the
pha
se
shift a
nd am
plitude
attenuatio
n
probl
em
[1
4]-[16].
This
pa
per sim
u
late
s comm
utation
failure
and
DC line
sh
o
r
t ci
rcuit faul
t unde
r
diffe
rent fault
co
ndition
s ba
sed o
n
Yun
n
an-
Guan
gdo
ng
80
0
k
V
UHV
DC t
r
a
n
smi
ssi
on
system in
PS
CAD/EMT
D
C the elect
r
o
m
agneti
c
simulatio
n
software. Th
e
aerial mo
d
e
com
pone
nt of the DC
curre
n
t tran
sient sign
al
d
I
is
decompo
se
d
into
six scale
s
of di
vision
s
with
the mo
dified mo
rp
holo
g
ical
filter.
The
morp
holo
g
ica
l
characte
risti
cs u
nde
r different
freq
uen
cies of the aerial mode com
pone
nt of
d
I
in
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 12, No. 4, Dece
mb
er 201
4: 753
– 762
754
the two
ki
n
d
s
of faults are
stu
d
ied
,
whi
c
h
will
give the t
heory to
di
stinguish bet
wee
n
comm
utation
failure an
d DC line short ci
rcuit fault.
2. Mathema
t
i
cal Morphol
og
y
Basic T
h
eor
y
2.1. The Basi
c Opera
t
ions
of Math
ema
t
ical Morpho
log
y
Erosi
on an
d
dilation are
the two ba
sic m
o
rp
holo
g
ical o
p
e
r
ations in m
a
th
ematical
morp
holo
g
y Corro
s
io
n an
d expan
sion.
The ope
rations a
r
e eq
ui
valent to filte
r
minimum a
nd
maximum va
lue withi
n
structu
r
in
g ele
m
ent of o
n
e
dime
nsi
o
n
a
l discrete
sign
al. In o
ne-
dimen
s
ion
a
l signal
s, ero
s
io
n is u
s
ed to suppress
the p
o
sitive pul
se
sign
al and
ke
ep the ne
gative
pulse signal. In
co
ntra
st
，
dilation i
s
u
s
ed to
su
ppre
s
s the
negati
v
e pul
se
sign
al an
d keep
the
positive pul
se
si
gnal. Let
f
(n)
a
s
o
n
e
-
dim
ensi
onal
sam
p
ling
sig
nal,
g
(n)
as one
dime
nsio
nal
stru
cturi
ng el
ement,
D
f
is the
definition do
main of
f
(n
)
and
g
D
is the d
e
finition domai
n of
g
(n)
.
The sig
nal o
f
the erosion
and the dilation com
put
ing formula
with stru
ctu
r
al element
s are
respe
c
tively shown in form
ula (1
) and fo
rmula (2).
(
f
g
)
(n)
m
in
[f
(n
m)
g(
m)]
(1)
(f
g
)
(
n
)
m
a
x
[
f
(
n
m
)
g
(
m
)
]
(2)
Whe
r
e,
f
nm
D
，
an
d
g
mD
.
It can
cre
a
te
a variety of
ope
ration
s t
o
co
mbin
e corrosi
on a
n
d
dilation. Th
e
openi
ng
operation i
s
t
hat first to
co
mpute e
r
o
s
io
n an
d the
n
to
dilation,
as shown in
form
ula (3). A
s
well
as, the clo
s
in
g operation is that
first to compute dilati
on and then t
o
ero
s
ion, a
s
sho
w
n in formula
(4).
(
f
g
)
(
n
)
(
(f
g)
g)(
n
)
(3)
(
f
g
)
(n
)
(
(f
g
)
g
)
(n
)
(4)
2.2. The Selection of Str
u
ctural Eleme
n
t
The shap
e a
nd len
g
th of
stru
cturi
ng el
ement
have
great affe
ct o
n
the re
sult
s
of sign
al
pro
c
e
ssi
ng [1
7]. Neverthel
ess, Similar to wavelet
an
alysis sele
cting basi
s
fun
c
tion, there is no
spe
c
ial
pri
n
ci
ple that
ca
n
be o
beyed t
o
when
se
l
e
ction pro
p
e
r
sha
pe
fo
r structuri
ng ele
m
ent
whi
c
h is. Th
a
t
is what kin
d
of structu
r
e
element
s
is
suitable for p
r
oce
s
sing the
sign
al, it needs to
test and prov
e again an
d again [18]. Linear, flat st
ru
cturin
g eleme
n
ts, squ
a
re, round, se
mici
rcle,
etc are often
used to pro
c
e
ss the p
o
wer syste
m
el
ectri
c
al si
gna
ls. And the more
compl
e
x the
sha
pe i
s
, the
greate
r
th
e a
m
ount of
co
m
putation
w
ill b
e
. The l
ength
of the
stru
ct
uring
elem
ent
is
importa
nt in
determi
ning
the filteri
ng
ch
ara
c
teri
st
ics.
In gen
eral, th
e sho
r
t st
ru
cturing
ele
m
ent
i
s
use
d
to capt
ure hi
gh freq
uen
cy and th
e morphol
ogi
ca
l filter i
s
u
s
ed to filter ou
t high freq
ue
ncy
noise
an
d sm
ooth signal. L
ong stru
ctu
r
in
g
elem
ent
ha
s g
ood l
o
w-p
a
ss p
e
rfo
r
ma
nce,
but it takes
longe
r time. Therefore, in
determinin
g
the s
hape
and length o
f
structu
r
ing
element, sig
nal
feature, targ
e
t
and comp
utational compl
e
xity should a
ll be taken int
o
accou
n
t.
2.3. Multi-Sc
ale Morphol
ogical Opera
t
ions
Multi-scal
e m
o
rph
o
logi
cal
operation i
s
a ro
ugh
to
fin
e
sig
nal
hierarchical p
r
o
c
ess to d
o
several e
r
o
s
i
on o
r
dilatio
n
ope
ration
s o
n
sig
nal u
s
in
g structu
r
e
elements of dif
f
erent
size. F
o
r
sign
al
f
(n)
and t
he
stru
ctural
element
g
(n)
, if
F
is the
mo
rph
o
logi
cal tran
sform, the m
u
l
t
i-scale
morp
holo
g
ica
l
operatio
n {
0
i
F|
i
} based on F can be defin
ed
as in formul
a
(5).
i
F
(f
)
i
F
(
f
/
i
)
(5)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Spectrum
Com
parative Study of Com
m
utati
on Failure and Short-Circuit .... (Chen Shi-long)
755
By the sam
e
token, th
e
multi-scal
e o
penin
g
an
d closin
g op
erati
ons
are a
s
shown in
formula (6) a
nd formul
a (7
).
(f
i
g
)(n
)
((f
i
g
)
i
g)(n)
(6)
(f
i
g
)(
n
)
(
(
f
i
g)
i
g
)
(
n)
(7)
2.4. The Con
s
truc
tion of
A Morpholo
g
ical Filter
Signal de
co
mpositio
n is
actually a p
r
oce
s
s
of mu
lti filtering on origi
nal si
gnal. Th
e
openi
ng a
n
d
the cl
osi
ng
operation
of morp
holo
g
y
have lo
w-pa
ss characte
ri
stic
and
ca
n
be
ca
scade
d to
comp
ose op
e
n
-cl
o
si
ng filte
r
an
d cl
ose-o
penin
g
filter.
Becau
s
e
the
output am
plitude
of open
-cl
o
si
ng filter is l
a
rger a
nd the
o
u
tput amplitu
de of cl
ose-o
penin
g
filter i
s
smalle
r, averag
e
combi
nation
of two kind
s o
f
filter is usual
ly adopted a
s
sho
w
n in [19
]
and formula
(8).
2
i
h
[
(f
g
g
)
(
n
)
(f
g
g
)
(
n)]
/
(8)
In the filtering
pro
c
ess, stru
cturin
g elem
e
n
ts
are like filter win
d
o
w
s.
Only whe
n
the sha
pe
of stru
ctural
element
s is
match
with the sh
ape of
sign
al, will the origin
al si
g
nal be retain
ed or
extracted. An
d whe
n
the type of st
ru
ctu
r
al element i
s
the sam
e
an
d the si
ze differs, the filteri
n
g
effec
t
is
not the s
a
me.
In this pa
per,
the wave
sh
a
pe of ele
c
tri
c
sign
al is to
o complex to be
analyzed u
s
in
g filter
sho
w
n
in fo
rmula
(8). So,
co
nsi
deri
ng
the ele
c
tri
c
al
sig
nal
cha
r
a
c
teri
stic, the
pape
r u
s
e
s
t
he
improve
d
filter sho
w
n in fo
rmula (9
) inste
ad.
11
2
3
24
5
6
7
12
17
y
2
12
ii
i
i
ii
i
i
i
ii
i
i
yf
g
g
g
g
f
gg
g
g
hy
y
/
i,
,
,
N
Gg
,
g
,
,
g
(9)
Whe
r
e,
1
y
is the
re
sult of
ope
n-cl
osi
ng
ope
ration;
2
y
is th
e
re
sult of
clo
s
e-op
enin
g
o
p
eration;
i
G
is assem
b
le o
f
structu
r
al el
ement of i-th scale.
2.5. Multi-Sc
ale Morphol
ogical Dec
o
mposition Al
gorithm
If
i
d(
n
)
is th
e de
compo
s
ition
shape
of the i
-
th
scal
e si
g
nal, multi-sca
l
e morphol
og
ical
algorith
m
s a
r
e as sho
w
n in
formula (1
0)
and form
ula (11).
11
11
1
12
ii
i
NN
(n)
(n)
(
df
(
n
)
h
dh
h
dh
i
n)
N
(10)
1
N
i
i
f
(
n)
d
(
n)
(11)
3. Spectrum
Analy
s
is of The Fault
Ba
sed
on
The Multi-Scale Morphology
Decomp
ositi
on
3.1. UHVDC
Simulation Model
This p
ape
r
use
s
PSCA
D
/EMTDC to
cre
a
te a simulation mo
del ba
sed
Yunnan
-
Guan
gdo
ng
80
0
k
V
UHV
DC tran
smissio
n
sy
stem as
sho
w
n
in Figure 1.
Its nominal v
o
ltage is
800
kV. Nomi
nal
curre
n
t is 3.125
kA. Tot
a
l in
stall
ed
capa
city is
50
00MW an
d th
e total le
ngth
of
transmissio
n line is 14
18 km.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 12, No. 4, Dece
mb
er 201
4: 753
– 762
756
Figure 1.
80
0
k
V
sim
u
lation mod
e
l
of Yunnan-G
uang
dong
UHVDC syste
m
Simulate DC line sh
ort ci
rcuit fault and commutation f
a
ilure b
a
sed
on the syste
m
sho
w
n
in Figure 1.
The foll
owing failures are
mainly
con
s
i
dere
d
.
Dire
ct
groun
ding
short
circuit
fa
ult
happ
ening i
n
the positive transmi
ssion li
nes
with faul
t point located
in the re
ctifier side, midp
oi
nt,
the inverte
r
side resp
ectiv
e
ly; Commut
a
tion failure f
ault is
cau
s
e
d
by sin
g
le
-p
hase g
r
oun
di
ng
fault with fault switchi
ng a
ngle of 270
° in the in
verter side AC
syst
em; Comm
utation failure f
ault
is ca
used by pha
se to pha
se shortin
g
fault with f
ault
swit
chin
g ang
le of 270° in the inverte
r
si
de
AC syste
m
; Comm
utation
failure fault i
s
cau
s
ed by
three p
h
a
s
e
grou
ndin
g
fa
ult in the inverter
side
AC
sy
ste
m
; Com
m
utat
ion failu
re i
s
caused
by t
he
large
r
tran
sfo
r
mation
ratio
of inverte
r
sid
e
conve
r
ter tra
n
sformer.
In this arti
cle,
the
sel
e
cte
d
el
ec
tr
ic
pa
r
a
me
te
r
is DC
c
u
rr
en
t
d
I
(p.u.)o
f
the inve
rter
side.
d
I
is de
com
p
o
s
ed by mode
decompo
sitio
n
. The ae
rial
mode
comp
o
nent of
d
I
is th
e analyzed
data. The sa
mpling fre
que
ncy is 20
0KHz.The
samp
li
ng time lengt
h is 100m
s (5
periodi
c waves).
3.2. Morphol
ogical Dec
o
mposition Steps
The a
e
rial
mode
com
p
o
nent of
DC
curre
n
t und
e
r
all
kind
s
o
f
fault con
d
itions i
s
decompo
se
d
by morp
holog
ical d
e
compo
s
ition. Power
system of el
e
c
tri
c
sig
nal h
a
s
a lot of noi
se
whi
c
h has a great deal of uncertainty.
T
he amplitude
of flat structu
r
i
ng element is 0. Its structure
is sim
p
le an
d
will not modi
fy the amplitude of orig
i
nal
signal. It is
more a
c
cu
rat
e
ly to extract the
morp
holo
g
ica
l
ch
aracte
risti
c
s of
sign
al
comp
ared wit
h
the structu
r
e of
no
n-ze
ro
ele
m
ent
[20].
So, ch
oo
sing
flat st
ru
cturi
ng el
eme
n
ts
on
sign
al
pr
oc
es
s is a g
ood
me
tho
d
to
g
e
t
id
ea
l filte
r
in
g
effect. Choo
se flat
stru
cturing el
ement
1
g
[0 0
0] a
s
the
unit
of st
ru
cturing
ele
m
ent
. Cal
c
ulat
e
stru
cturi
ng el
ement a
s
sem
b
les
of vario
u
s
scal
es a
c
cordin
g to fo
rmula
(1
)
k
1
Rr
[21]. The
step
s of
6
scales mo
rph
o
logical de
co
m
positio
n
of th
e ae
rial
mod
e
compo
nent
of DC
cu
rre
n
t
sign
al
d
I
unde
r different fault con
d
ition
s
are as follo
ws.
(1)
De
comp
ose
inverter side DC cu
rre
nt
d
I
(p.u.) ba
sed o
n
morp
holo
g
i
c
al de
com
p
o
s
ition. And
extract the ae
rial mod
e
co
mpone
nt of
d
I
.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Spectrum
Com
parative Study of Com
m
utati
on Failure and Short-Circuit .... (Chen Shi-long)
757
(2)
Cal
c
ulate
stru
cture el
eme
n
t assemble u
n
der different scale
s
.
(3)
Acco
rdi
ng to
formula
(10),
cal
c
ulate
filtering
output
i
h
of the a
e
rial
m
ode
co
mpon
e
n
t of
DC
cur
r
e
n
t
sign
al
d
I
under different scale
s
.
(4)
Acco
rdi
ng to
step (2
), calc
ul
ate filtering output u
nder
scale
i
=
1
,2,3, ...,5.
The detail
comp
one
nt u
nder different
scale
1
d
~
5
d
an
d
the ap
proxim
ation
6
d
of the
sixth scal
e
ca
n b
e
cal
c
ulate
d
according to formula (1
0).
3.3. Multi-Sc
ale Morphol
ogical Dec
o
mposition Spectr
u
m Ana
l
y
s
is Form
DC lin
e sho
r
t circuit fault h
appe
ns o
n
transmi
ssion li
ne at 300
km,
700 km, 120
0 km at
0.3s an
d last
s for 100
ms.
T
he morphol
ogical decom
positio
ns of the aeri
a
l mo
de com
pon
e
n
t of
d
I
unde
r differe
nt scale
s
are
sho
w
n in
Fi
gure
2(a)
,(b),
(
c). Three p
h
a
se
gro
undi
n
g
fault in the
inverter
sid
e
AC
syste
m
hap
pen
s at 0.3s
a
nd la
sts fo
r 100m
s. Th
e morphol
og
ical
decompo
sitio
n
s
of the a
e
ri
al mod
e
com
pone
nt of
d
I
un
der
differe
nt
scale
s
a
r
e
sh
own
in 3
(
a
)
.
The pha
se to
phase
sho
r
ting fault with fault sw
itchi
n
g angle of 27
0 ° happe
ns
on the inverter
side A
C
syst
em and la
sts for 100m
s.T
he morphol
o
g
ical d
e
comp
osition
s
of th
e aeri
a
l mod
e
comp
one
nt of
d
I
und
er
different scale
s
are shown in
3(
b). Th
e si
ngle
pha
se
gro
u
n
d
ing fault
with
swit
chin
g an
gle of 270 °
happ
en
s on
the invert
er
side A
C
syst
em and la
sts for 100m
s.T
h
e
morp
holo
g
ica
l
deco
m
po
sitions of the a
e
r
ial mod
e
co
mpone
nt of
d
I
unde
r differe
nt scal
e
s
are
sho
w
n in 3
(
c). Increa
se the
transfo
rmatio
n ratio of
the inverter
side t
r
an
sform
e
r to
3.5 that leads
to the com
m
u
t
ation failure.
The mo
rph
o
l
ogical de
com
positio
ns
of the ae
rial mo
d
e
com
pon
ent
of
d
I
under diffe
re
nt scal
e
s a
r
e
sho
w
n in Fig
u
re 3
(
d).
0
0.
5
1
1.
5
2
x 1
0
4
0
1
2
0
0.
5
1
1.
5
2
x 1
0
4
0
1
2
0
0.
5
1
1.
5
2
x 1
0
4
-
0.2
0
0.2
0
0.
5
1
1.
5
2
x 1
0
4
-
0.2
0
0.2
0
0.
5
1
1.
5
2
x 1
0
4
-
0.2
0
0.2
0
0.
5
1
1.
5
2
x 1
0
4
-
0.2
0
0.2
0
0.
5
1
1.
5
2
x 1
0
4
-1
0
1
Dat
a N
umb
er
0
0.5
1
1.5
2
x 10
4
0
1
2
0
0.5
1
1.5
2
x 10
4
0
1
2
0
0.5
1
1.5
2
x 10
4
-0.5
0
0.5
0
0.5
1
1.5
2
x 10
4
-0.5
0
0.5
0
0.5
1
1.5
2
x 10
4
-0.5
0
0.5
0
0.5
1
1.5
2
x 10
4
-0.5
0
0.5
0
0.5
1
1.5
2
x 10
4
-1
0
1
Dat
a Numbe
r
(a)
DC lin
e dire
ct
groun
ding
sh
ort
(b) DC line dire
ct groundi
ng short
circuit fault at 300km
circuit fault at 700km
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 12, No. 4, Dece
mb
er 201
4: 753
– 762
758
0
0.
5
1
1.
5
2
x 1
0
4
0
1
2
0
0.
5
1
1.
5
2
x 1
0
4
0
1
2
0
0.
5
1
1.
5
2
x 1
0
4
-
0.5
0
0.5
0
0.
5
1
1.
5
2
x 1
0
4
-
0.5
0
0.5
0
0.
5
1
1.
5
2
x 1
0
4
-
0.5
0
0.5
0
0.
5
1
1.
5
2
x 1
0
4
-
0.5
0
0.5
0
0.
5
1
1.
5
2
x 1
0
4
-1
0
1
Dat
a N
umb
er
(c)
DC lin
e dire
ct
groun
ding
sh
ort circuit faul
t at 1200km
Figure 2.
Multi-scale m
o
rp
h
o
logy de
com
positio
n for the aerial m
ode
compo
nent o
f
d
I
of DC
line sh
ort ci
rcuit faults
0
0.
5
1
1.5
2
x 1
0
4
0
3
0
0.
5
1
1.5
2
x 1
0
4
0
3
0
0.
5
1
1.5
2
x 1
0
4
-0.
05
0
0.
05
0
0.
5
1
1.5
2
x 1
0
4
-0.
05
0
0.
05
0
0.
5
1
1.5
2
x 1
0
4
-0.
05
0
0.
05
0
0.
5
1
1.5
2
x 1
0
4
-0.
05
0
0.
05
0
0.
5
1
1.5
2
x 1
0
4
-0
.2
0
0.
2
Dat
a N
um
ber
0
0.
5
1
1.
5
2
x 10
4
0
3
0
0.
5
1
1.
5
2
x 10
4
0
3
0
0.
5
1
1.
5
2
x 10
4
-0.05
0
0.05
0
0.
5
1
1.
5
2
x 10
4
-0.05
0
0.05
0
0.
5
1
1.
5
2
x 10
4
-0.05
0
0.05
0
0.
5
1
1.
5
2
x 10
4
-0.05
0
0.05
0
0.
5
1
1.
5
2
x 10
4
-0.2
0
0.
2
Da
ta N
umbe
r
(a)
Thre
e pha
se
grou
ndin
g
fau
l
t in the
(b) Phase to pha
se
shorting fault i
n
inverter s
i
de AC s
y
s
t
em
the inverter s
i
de AC
s
y
s
t
em
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Spectrum
Com
parative Study of Com
m
utati
on Failure and Short-Circuit .... (Chen Shi-long)
759
0
0.
5
1
1.
5
2
x 10
4
0
3
0
0.
5
1
1.
5
2
x 10
4
0
3
0
0.
5
1
1.
5
2
x 10
4
-
0.05
0
0.05
0
0.
5
1
1.
5
2
x 10
4
-
0.05
0
0.05
0
0.
5
1
1.
5
2
x 10
4
-
0.05
0
0.05
0
0.
5
1
1.
5
2
x 10
4
-
0.05
0
0.05
0
0.
5
1
1.
5
2
x 10
4
-0.2
0
0.2
Data N
umber
0
0.
5
1
1.
5
2
x 10
4
0
1
2
0
0.
5
1
1.
5
2
x 10
4
0
1
2
0
0.
5
1
1.
5
2
x 10
4
-0.0
5
0
0.0
5
0
0.
5
1
1.
5
2
x 10
4
-0.0
5
0
0.0
5
0
0.
5
1
1.
5
2
x 10
4
-0.0
5
0
0.0
5
0
0.
5
1
1.
5
2
x 10
4
-0.0
5
0
0.0
5
0
0.
5
1
1.
5
2
x 10
4
-0.
2
0
0.
2
Data
Num
ber
(c)
Single-pha
se
grou
ndin
g
fau
l
t in the
(d) Inverter
s
i
de
c
o
n
v
er
te
r
tr
an
s
f
or
me
r
inverter s
i
de AC s
y
s
t
em
ratio of 3.5
Figure 3. Multi-scale m
o
rp
h
o
logy de
com
pos
itio
n for the aerial m
ode
compo
nent o
f
d
I
of
comm
utation failure
As i
s
sho
w
n
in Fi
gure 2
and
Figu
re
3, un
de
r di
fferent fault
con
d
ition
s
, the hi
g
h
freque
ncy
co
mpone
nt on
the
same
sca
l
e of the
aeri
a
l mod
e
com
pone
nt of
d
I
in the DC li
ne
sho
r
t circuit f
ault and the
comm
utation
failure
h
a
s dif
f
erent chan
gi
ng tren
ds.
When the
DC li
ne
sho
r
t circuit f
ault occu
rs in
different po
sition,
the high
freque
ncy compon
ent of the aeri
a
l mo
de
comp
one
nt of
d
I
durin
g the fault duration
is rich, an
d
grad
ually de
crea
se
s. Furth
e
rmo
r
e, the
more th
e DC line short
circuit cl
ose to
the inve
rte
r
, the greate
r
th
e ch
ang
e fre
quen
cy of hi
gh
freque
ncy compon
ent, and the bigg
er the ampli
t
ude increa
ses. In Figure 3, when
the
comm
utation
failure o
c
curs cau
s
ed by
different
fault
con
d
itions, t
he chang
e chara
c
te
risti
c
s of
the high freq
uen
cy com
p
o
nent
1
d
~
5
d
of the
aerial
mode
compon
ent of
d
I
is very different with
of the DC line short
circui
t fault.. Whereas i
n
Figure 3(a),(b),(c),(d)
, the low frequency
change
cha
r
a
c
t
e
ri
st
ic
s in
6
d
of the aerial m
ode
comp
one
nt of
d
I
all are o
b
vious. Th
e mo
rphol
ogi
cal
spe
c
tru
m
of
the sixth
sca
l
e is
analy
z
e
d
u
s
ing
Fou
r
ier tran
sform.
As i
s
sho
w
n in Fi
gure
4,
0Hz~100
Hz in freque
ncy range in
6
d
of the aerial mo
de
comp
one
nt of
d
I
is richer.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 12, No. 4, Dece
mb
er 201
4: 753
– 762
760
(a)
Thre
e pha
se
grou
ndin
g
fau
l
t in the inverter sid
e
AC sy
stem
(b)
Phase to ph
a
s
e shortin
g
faul
t in the inverter si
de AC
system
(c) Single ph
ase g
r
ou
ndin
g
fault in the inverter
side
AC system
0
10
0
20
0
300
40
0
50
0
0
0.
5
1
1.
5
2
2.
5
x 1
0
4
F
r
eq
uen
c
y
/
H
z
A
m
p
lit
u
d
e
0
10
0
20
0
30
0
40
0
50
0
0
0.
5
1
1.
5
2
2.
5
x 1
0
4
F
r
eq
ue
nc
y
/
H
z
A
m
p
lit
u
d
e
0
10
0
20
0
30
0
40
0
500
0
0.
5
1
1.
5
2
2.
5
x 1
0
4
F
r
e
quen
c
y
/
H
z
A
m
p
lit
u
d
e
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Spectrum
Com
parative Study of Com
m
utati
on Failure and Short-Circuit .... (Chen Shi-long)
761
(d) Inve
rter si
de co
nverte
r tran
sform
e
r ra
tio of 3.5
Figure 4. FFT
spe
c
trum of
sixth scale m
o
rph
o
logi
cal
spe
c
tru
m
of the aeri
a
l mod
e
comp
one
nt of
d
I
4. Conclusio
n
This pap
er simulates co
mmutation fa
ilure
and short
c
i
rc
uit fault in PSCA
D
/EMTDC
based on the
model of Yunnan
-Gu
ang
zhou
80
0
k
V
UHVDC transmi
ssion
system. The
fault
points of DC line groun
di
ng sho
r
t circuit faults
are
located at 300km, 70
0km, 1200km o
f
the
positive tran
smissi
on line
s
. Simulate commutation f
a
ilure
s cau
s
e
d
by three p
hase gro
undi
ng
fault, single
-
p
hase groundi
ng fault, pha
se to pha
se sh
orting fault in
the inverter
si
de AC sy
ste
m
and i
n
crea
si
ng tra
n
sfo
r
m
a
tion ratio of
inverte
r
si
d
e
tran
sfo
r
me
r to 3.5. Extract the
tran
sient
sign
al of inverter si
de DC.
Then de
com
pose the aeri
a
l mode com
pone
nt of
d
I
into 6 different
scale
s
. At last, analyze the
high freq
uen
cy feature
s
of
1
d
~
5
d
and lo
w freque
ncy feat
ure
s
of
6
d
in the two kinds of fault case
s. The anal
ysis re
sults show that the
aerial mod
e
compon
ent of
ܫ
ௗ
has m
o
re
hig
h
frequ
en
cy comp
one
nt in
1
d
~
5
d
in the ca
se of sho
r
t ci
rcuit fault. While in the
ca
se of co
m
m
utation failu
re the ae
rial
m
ode comp
o
nent ha
s less frequen
cy di
vision
s in
6
d
.
Referen
ces
[1]
Yuan
Qin
g
y
un.
Prese
n
t state
and
a
ppl
icatio
n
pros
pect
of ul
tra HVD
C
tran
smission
i
n
C
h
ina.
Po
wer
System
Technology.
200
5; 29(14): 1-3.
[2]
Lin L
i
ng
xue, Z
H
ANG Yao, et al. A surve
y
on
co
mmutation f
a
ilur
e
in mu
lti-i
n
feed HV
DC transmissi
o
n
s
y
stems.
Pow
e
r System T
e
ch
nol
ogy.
20
06; 30(1
7
): 40-4
6
.
[3]
Liu Ji
an, L
i
Xi
n
g
y
ua
n, W
u
Ch
ong, et a
l
. Res
earch
on critic
al in
de
x
of co
mmutation fa
il
ure i
n
HVD
C
sy
s
t
e
m
.
Power
System
Technology.
200
9; 33(8): 8-12.
[4]
W
ang Ga
ng, L
i
Z
h
ike
ng, H
u
a
ng Mi
n, et a
l
.
Influenc
e of
ini
t
ial fa
ult vo
ltag
e a
ngl
e o
n
co
mmutati
o
n
failure identifi
c
a
tion
in
a
HV
D
C
s
y
stem.
Aut
o
mation
of Electric Power Sy
stem
s.
201
0; 3
4
(4): 4
9
-5
4,
102.
[5]
Che
n
Shu
y
o
n
g
,
Li Xin
n
i
an, Yu Jun, et al.
A metho
d
bas
ed on the si
n-
cos compo
nen
ts detectio
n
miti
gates co
mmutati
on fai
l
ur
e in HVD
C
. Proceedings of the C
SEE. 2005; 25(14): 1-6.
[6]
He Ch
aor
on
g, Li
Xi
ng
yu
an, Ji
n Xia
o
min
g
, et
al. Criteri
a
for c
o
mmutatio
n
fail
ure i
n
HVD
C
transmissi
o
n
s
y
stems.
Pow
e
r System T
e
ch
nol
ogy.
20
06; 30(2
2
): 19-2
4
.
[7]
He Cha
o
ro
ng, Li Xin
g
y
ua
n, Jin Xi
aomi
ng, et al. Si
mulati
on
ana
l
y
sis on co
mmutation fai
l
u
r
e criteria fo
r
HVDC transmi
ssion s
y
stems.
Pow
e
r System T
e
chnol
ogy.
2
007; 31(
1): 20-
24.
[8]
Xu So
ng
lin, Hu
ang Sh
ao-
xi
an.
F
ault diag
nosi
s
of commutation fail
ures in th
e HVDC s
y
ste
m
based o
n
the
w
a
vel
e
t e
n
e
rg
y s
pectrum
and
gre
y
comp
rehe
nsive
rel
a
ti
onsh
i
p
de
gree.
Power System Protection
and C
ontrol.
2
012; 40(
3): 85-
89.
0
100
200
300
400
500
0
0.
5
1
1.
5
2
2.
5
x 1
0
4
F
r
equenc
y
/
H
z
Am
p
lit
u
d
e
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 12, No. 4, Dece
mb
er 201
4: 753
– 762
762
[9]
Lin L
i
ng
xue, Z
han
g Yao, Z
h
ong Qin
g
, et al.
F
ault iag
n
o
s
is of commutation fai
l
ur
es i
n
the HVD
C
system b
a
sed
on a
metho
d
of
w
a
velet ener
g
y
statistics
.
[10]
Z
hang
Yao,
Li
n Li
ng
xue, Z
h
o
ng Qin
g
. F
a
u
l
t
dia
gnos
is b
a
se
d on
w
a
v
e
l
e
t for commutati
o
n
fail
ures
i
n
HVDC.
J
ourn
a
l
of S
outh
Ch
i
na
Univ
ersity
of
T
e
chn
o
l
ogy
(Natur
al Sc
ie
nce E
d
itio
n).
2
007;
35(
10):
172-
177.
[11]
Yin W
enq
in, L
i
u Qian
jin. Ma
thematica
l
mor
pho
log
y
rev
i
e
w
a
nd its ap
pl
icatio
ns in p
o
w
e
r s
y
stem.
RELAY.
200
7; 35(1
9
): 172-
17
7.
[12]
Shu
Hon
g
ch
un
, W
ang
Cha
o
,
Z
hang
Jie
W
u
Na. HVD
C
tra
n
smissio
n
s
y
st
em fau
l
t id
entif
icatio
n a
n
d
locati
ng
alg
o
rit
h
m usi
n
g
mat
hematic
al m
o
r
pho
log
y
.
El
e
c
tri
c
Po
we
r Au
to
ma
ti
on
Equ
i
pm
en
t.
20
07;
27(4): 6-9, 18.
[13]
Shu H
o
n
g
chu
n
, Che
ng
Ch
unh
e, Z
hao
W
e
n
y
ua
n
，
Zang Jiabei. Accurate
trav
eli
n
g
-
w
a
ve fro
n
t
detectio
n
b
a
se
d o
n
mor
pho
lo
g
y
a
nd
HHT
.
Electric Pow
e
r
Auto
mati
on E
qui
p
m
ent.
2
0
0
9
; 29(
7): 1-
7
,
37.
[14]
Z
hao Ju
n, Lv Yanp
ing, W
a
n
g
Han
g
u
a
n
g
. Ne
w
sch
eme t
o
ide
n
tif
y
lig
ht
nin
g
distur
ban
ce for the UHV
transmissio
n
li
nes bas
ed on
multi
2
scal
e
morph
o
lo
g
y
d
e
comp
ositio
n.
High
Vo
ltag
e
Engi
neer
in
g.
200
9; 35(5): 99
4-99
8.
[15]
W
ang Ji
ng, S
hu H
o
n
g
-chu
n, Che
n
Xu
e-
yu
n.
Multi-sca
le
mor
p
h
o
lo
gy a
nalysis
of dy
n
a
mic p
o
w
e
r
qua
lity disturb
a
n
ces
. Procee
di
ngs of the CSE
E
. 2004; 24(
4): 63-67.
[16]
Z
eng Ji
yo
ng, D
i
ng Ho
ngfa, Du
an Xia
n
zho
ng.
Har
m
on
ics det
ection a
nd dist
urba
nce loc
a
tio
n
meth
od
s
base
d
on
math
ematica
l
morp
hol
ogy
. Proce
e
d
in
gs of the C
SEE. 2005; 25(
11): 57-6
2
.
[17]
Ou
yang Se
n, Hua
ng Ru
nh
on
g. A method of
locatin
g
transi
ent disturb
anc
e of po
w
e
r q
u
a
lit
y
b
a
se
d o
n
morph
o
lo
gica
l edg
e detecti
on
.
Power Syste
m
Technology.
2012; 3
6
(4): 6
3
-67.
[18]
Z
hang
Li
ju
n, Y
ang
De
bi
n,
Xu
Jin
w
u,
Ch
en Z
h
i
x
in.
Appr
oac
h to
e
x
tracting
gear
fault fe
atu
r
e b
a
sed
o
n
mathematic
al morph
o
lo
gica
l filterin
g.
Chi
n
e
s
e Jour
na
l of
Mecha
n
ica
l
En
gin
eeri
ng.
2
0
0
7
; 43(
2): 71
-
75.
[19]
Che
n
Pan, C
h
en Ha
o
y
o
ng,
Ye Ron
g
. W
i
n
d
spee
d forec
a
sting b
a
se
d o
n
multi-sca
le
morph
o
lo
gica
l
analy
sis.
Pow
e
r System Prote
c
tion an
d Co
ntrol.
201
0; 38(2
1
)
: 12-18.
[20]
Li T
i
an
yu
n, Ya
ng M
e
i, Z
h
o
u
Xic
h
a
o
, et
al.
Met
hod
of par
tial disch
arge
sign
al an
al
ysis
bas
ed o
n
w
a
vel
e
t transfo
rm and mathe
m
atical mor
p
h
o
lo
g
y
.
Pow
e
r S
ystem T
e
ch
nol
ogy
. 200
7; 31(
6): 56-60
．
[21]
Z
hu Sh
ihu, Z
h
u H
ong,
He P
e
izh
ong.
Res
e
arch
on
struct
urin
g e
l
eme
n
ts selecti
o
n
in
morph
o
lo
g
y
oper
ation.
Mo
d
e
rn Co
mputer
. 200
9; 3(1): 19–
21.
Evaluation Warning : The document was created with Spire.PDF for Python.