Indonesi
an
Journa
l
of El
ect
ri
cal Engineer
ing
an
d
Comp
ut
er
Scie
nce
Vo
l.
23
,
No.
3
,
Septem
ber
20
21
,
pp.
13
31
~
133
9
IS
S
N: 25
02
-
4752, DO
I: 10
.11
591/ijeecs
.v
23
.i
3
.
pp
1331
-
133
9
1331
Journ
al h
om
e
page
:
http:
//
ij
eecs.i
aesc
or
e.c
om
Modelin
g of
h
i
gh frequen
c
y hi
gh v
oltage
of wav
eform
s o
n
life of
e
n
amel ins
ulation
S
.
Na
r
as
im
h
a Rao,
El
an
ser
al
at
h
an
K
as
ina
th
an
Depa
rtment
o
f
E
le
c
tri
c
al a
nd
Ele
ct
roni
cs
Engi
n
eering
,
Puduch
err
y
Technol
ogi
cal U
nive
rsit
y
,
Puducher
r
y
,
Indi
a
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Ma
y
21
,
2021
Re
vised
Ju
l
30
,
2021
Accepte
d
Aug
4
,
2021
In
rec
en
t y
ea
rs i
t
has
bee
n
observ
ed
tha
t insulatio
n
fai
lur
e
in
e
le
c
t
ric
a
l
m
otors
is
ca
used
b
y
ad
justa
ble
spee
d
drive
s
fed
b
y
p
ower
el
e
ct
ron
ic
conve
rt
ers.
The
se
conve
r
te
r
s
produc
e
impuls
e
wave
form
s
havi
ng
a
high
slew
ra
t
e
gene
ra
te
d
b
y
the
high
sw
it
chi
ng
fr
eque
nc
y
o
f
IGBTs.
Thi
s
pap
er
foc
uses
on
high
sw
it
chi
ng
fre
quency
str
ess
in
low
vo
lt
ag
e
el
e
ct
r
ic
a
l
m
otors
for
adj
ustable
spee
ds.
To
exa
m
in
e
the
m
otor
win
ding
insula
t
ion
under
such
stress
twiste
d
-
pa
ir
sam
ple
s
were
deve
lop
ed
from
ena
m
el
ed
wir
es.
A
single
-
coa
t
ed
p
ol
y
es
te
r
of
en
amel
wi
th
a
thi
ckn
ess
of
40
m
ic
rons
is
u
sed
for
th
is
work.
High
-
fre
q
uenc
i
es,
high
vo
lt
ag
es
of
Squar
e,
and
Squar
e
-
risi
ng,
Square
-
spike
wave
fo
rm
s
of
10
-
30
kHz
are
used
her
e.
T
he
t
est
r
esult
s
sh
ow
tha
t
the
insula
ti
on
f
ails
ea
rl
ie
r
for
th
e
Sq
uar
e
wave
for
m
compare
d
to
the
Square
-
spike
and
Squar
e
-
rising
wav
efo
r
m
s.
In
a
nutshell
,
th
ere
is
an
ana
l
y
sis
of
PD
form
at
ion
in
th
e
insula
t
ion
s
y
stem
at
a
highe
r
sw
it
chi
ng
fr
eque
nc
y
is
ana
l
y
z
ed
.
E
le
c
tr
ic
fi
el
d
d
istri
but
ions
bet
we
en
tw
iste
d
pa
irs
with
var
iou
s
air
gaps
of
the
insula
ti
on
s
y
s
te
m
stressed
b
y
the
Square
and
Square
-
risin
g
wave
form
s up
to
30
kHz
ar
e
m
odel
ed
using COMS
OL
software
.
Ke
yw
or
d
s
:
Ele
ct
ric fiel
d d
ist
ribu
ti
on
High s
witc
hing
f
re
quency
Mod
el
in
g
Square
-
risin
g wav
e
f
or
m
Twiste
d
-
pair
This
is an
open
acc
ess arti
cl
e
un
der
the
CC
B
Y
-
SA
l
ic
ense
.
Corres
pond
in
g
Aut
h
or
:
S. Narasim
ha
Ra
o
Dep
a
rtm
ent o
f El
ect
rical
an
d
Ele
ct
ro
nics
E
nginee
rin
g
Pudu
c
he
rr
y Te
chnolo
gical
Universit
y,
P
uduc
herry
Pil
la
ic
hav
adi,
Pudu
c
he
rr
y,
I
ndia
60
5014
Em
a
il
:
sn
rinskit
@p
ec.e
du
1.
INTROD
U
CTION
In
i
ndus
tria
l
dri
ves,
the
contr
ol
o
f
s
pee
d
is ac
com
plished
b
y
us
in
g power
elec
tro
nic conve
rters
, usi
ng
the
P
W
M
te
ch
nique.
T
his
te
chnolo
gy
bro
ught
ne
w
dev
el
op
m
ents
in
the
sp
eed
c
on
t
ro
l
of
in
dustria
l
m
oto
rs
[1
]
.
H
oweve
r,
these
co
nv
e
rters
ge
ner
at
e
hi
gh
switc
hi
ng
im
pu
lse
s
at
eve
ry
instant
of
s
witc
hing,
w
hic
h
is
of
ver
y
hi
gh
sle
w
rate
and
high
rep
et
it
ive
f
requen
cy
produce
d
by
I
GBT
’s.
Du
e
t
o
this,
th
e
enam
el
insu
l
at
ion
of
the
coil
is
sub
j
ect
ed
t
o
re
pe
at
ed
surge
vol
ta
ges
with
s
ho
rt
rise
tim
es
i
n
the
orde
r
of
nanosec
onds.
These
su
r
ges
in
duce
reli
abili
ty
pr
oblem
s
in
low
and
m
edium
rat
ing
s
of
th
e
m
otor
ins
ulat
ion
syst
e
m
[2
]
.
As
this
wav
e
f
or
m
trav
el
s
from
the
inv
e
rter
to
t
he
m
oto
r
e
nd,
t
he
s
pik
e
gets
enh
a
nce
d
i
n
m
agn
it
ud
e
du
e
to
t
he
m
is
m
a
tc
h
of
i
m
ped
ance
bet
ween
t
he
in
vert
er,
m
oto
r,
an
d
cable.
The
ph
eno
m
enon
of
r
eso
nan
ce
a
nd
r
eflect
ion
can ind
uce twi
ce the s
uppl
y v
oltage at t
he
tu
rn
-
to
-
tur
n
i
nsula
ti
on
of
sta
tor win
ding [3
]
,
[
4],
wh
ic
h
is t
he m
aj
or
reason for
the
un
e
xpect
ed
fail
ur
e of w
i
nd
i
ng o
f
insu
la
ti
on. Fur
t
her
m
or
e, t
he
distrib
utio
n
of
elec
tric
al
p
ot
entia
l
is
no
n
-
un
if
orm
acro
ss
the
wi
nd
i
ngs
[5
]
,
[
6].
Hen
ce
,
ab
nor
m
al
e
le
ct
rical
stress
is
exp
er
ie
nced
by
pha
se
to
ph
a
se
windin
g
of
m
oto
r
ins
ul
at
ion
,
m
or
e
stress
tha
n
sine
wav
e
f
or
m
[7
]
,
at
the
first
tu
rn
s
of
t
he
ins
ul
at
ion
syst
e
m
[8
]
,
in
par
ti
cula
r.
O
wing
to
a
n
overs
hoot
in
volt
age
peak
s,
th
e
par
t
ia
l
dischar
ge
ta
kes
place
in
wi
nd
i
ng
because
of
w
hich
the
li
fetim
e o
f
org
a
nic e
na
m
el
ed
insu
la
ti
on is
rem
ark
abl
y
sh
ort
[9
]
,
[
10]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
3
,
Se
ptem
ber
20
21
:
13
3
1
-
133
9
1332
Investi
gatio
n
of
lo
w
-
fr
e
quenc
y
stress
is
re
po
rted
in
the
li
te
r
at
ur
e.
H
ow
e
ve
r,
as
t
he
e
nam
e
l
insu
la
ti
on
m
ay
exp
erienc
e
hig
he
r
fr
e
qu
ency,
this
wor
k
pr
e
sents
hi
gher
f
reque
ncy
stress
of
10
-
30
kH
z
of
Square
,
Square
-
risin
g,
and
S
qu
a
re
-
sp
i
ke
wav
e
f
or
m
s.
Test
ing
with
Square
wa
vefo
rm
s
pr
oduc
es
highest
dv
/
dt
that
can
cause
high
str
ess
on
the
i
nsula
ti
on
w
hen
com
par
ed
with
oth
e
r
wa
ve
form
s.
The
reas
on
f
or
c
hoos
i
n
g
hi
gh
fr
e
qu
e
ncy
f
or
the
te
st
is
becau
se
of
the
gro
wing
dem
and
for
var
ia
ble
spe
ed
ap
plica
ti
on
s
an
d
the
inc
reasin
g
nu
m
ber
of
low
vo
lt
age
m
oto
r
s
[11].
The
pu
rpose
of
em
pl
oying
the
high
fr
eq
ue
ncy
swi
tc
hin
g
in
t
he
present
work
is
to
exa
m
ine
the
el
ect
ric
stress
pro
duced
by
t
he
hig
h
-
f
re
qu
e
ncy
pu
lse
s
with
th
e
sh
a
rp
rate
of
rise
of
vo
lt
age
s
urge
on the
m
oto
r wind
i
ng ins
ulati
on.
The
a
ging
of
enam
el
insu
la
te
d
wires
w
as
inv
est
ig
at
ed
usi
ng
fiv
e
diff
e
re
nt
wa
ve
form
s
of
a
fun
dam
ental
fr
equ
e
ncy
of
30
0
Hz
in
the
presence
of
P
Ds
and
re
porte
d
that
the
li
fe
of
enam
el
insu
la
ti
on
is
aff
ect
ed
by
the
sh
a
pe
of
the
wav
e
f
or
m
[1
2]
.
The
e
ff
ect
of
short
unip
olar
rep
et
it
ive
im
pu
lsi
ve
vo
lt
age
s
with
200
ns
volt
age
durati
on
on
P
D
sta
ti
sti
cs
an
d
in
su
la
ti
on
li
f
et
i
m
e
of
e
nam
el
ed
wi
res
wa
s
stu
di
ed
[
1
3
]
.
It
is
repor
te
d
that
PD
m
agn
it
ude
unde
r
un
i
po
l
ar
s
hort
im
pu
lsi
ve
volt
ages
with
diff
e
ren
t
rise
tim
es
increase
sign
ific
a
ntly
with
decr
easi
ng
rise
tim
es.
The
l
ife
of
two
ki
nd
s
of
enam
el
wires
of
orga
nic
resin
an
d
nano
-
fill
ed
resin
we
r
e
stud
ie
d
an
d
m
od
el
ing
of
a
li
fetim
e
was
predict
ed
[
1
4
]
.
I
F
Ra
dzi
et
al
.
[1
5
]
ar
gued
t
hat
there
is
an
i
m
pr
ov
e
m
ent
of
de
gr
a
da
ti
on
at
long
te
rm
by
PD
in
a
dr
ast
ic
way
w
he
n
the
ap
plied
vo
lt
age
is
6.5
kV
to
8.5
kV.
Brea
kdow
n
stu
dies
with
a
ste
p
of
50V/s
f
or
AC
and
a
ste
p
of
100
V/s
f
or
DC
on
t
hin
-
film
insu
la
ti
on
with
a
thick
ne
ss
of
100
μm
w
ere
done
[1
6
]
.
It
is
al
so
repor
te
d
that
the
br
ea
kdow
n
str
eng
t
h
of
com
po
sit
es
is
lowe
r
c
om
par
ed
to
un
fill
ed
poly
pro
pyle
ne.
Re
searche
rs
re
ported
that
the
fail
ur
e
of
t
wiste
d
pai
r
ins
ulat
ion
i
s
du
e
t
o
hi
gh
re
petit
ion
f
reque
ncy
init
ia
ti
ng
the
par
ti
al
dis
charge
[
17]
.
It
is
al
so
rep
or
t
ed
that
thr
ough
the
si
m
ulati
on
s the
PD arises i
n
t
he
m
ic
ro
gap b
e
tween t
he
twist
s u
si
ng the C
O
MSOL s
oft
ware.
Th
ough
the
re
a
re
works
done
at
a
lo
wer
f
requen
cy
up
t
o
60
0
Hz
a
nd
at
3
kH
z
,
t
he
li
fe
of
ins
ulati
on
su
bject
e
d
to
hig
he
r
fr
e
quency
pulse
s
due
t
o
M
OS
FET
or
I
GBT
s
witc
hing
of
up
to
20
-
40
kHz
are
ne
eded
t
o
be
kn
own.
The
sta
tor
wi
nd
i
ng
insu
la
ti
on
is
s
ubj
ect
ed
to
s
uc
h
hi
gh
fr
e
que
nc
y
stress
w
hich
is
a
m
or
e
rea
li
stic
on
e
.
As
ob
se
r
ve
d
by
resea
rchers
that
on
e
ca
nnot
ass
um
e
a
theo
reti
cal
deducti
on
that
lo
w
-
fr
eq
ue
ncy
te
st
data
can
be
us
e
d
to
pr
edict
li
fe
with
increasin
g
f
reque
ncy
fo
r
a
ll
insu
la
ti
on
syst
e
m
s,
the
te
st
was
done
at
10
kHz
[18]
bu
t
the
obj
ect
iv
e
was
to
save
the
po
wer
loss
rather
than
to
k
now
the
li
fe
of
the
insu
la
ti
on.
He
nce
a
n
at
tem
pt
is
m
ade
in
th
is
wo
r
k
to
te
st
the
twist
ed
pair
ins
ulati
o
n
with
the
higher
f
reque
nc
y
of
10
-
30
kHz
to
inv
est
igate
t
he l
ife o
f
twist
ed pair e
nam
el
insu
la
ti
on
.
In t
he prese
nt st
ud
y, t
he
twist
e
d
-
pair
sam
ple o
f po
l
ye
ste
r
enam
el
is
stressed
by
S
qu
a
re,
Square
-
risi
ng,
an
d
S
quare
-
spi
ke
wa
ve
form
s
over
t
he
range
of
f
reque
ncies
from
10
-
30
kHz.
Th
e
te
sts
are
perf
or
m
ed
to
s
how
that
early
fail
ur
e
of
twist
ed
pair
i
ns
ulati
on
co
uld
be
as
soc
ia
te
d
with
P
D
ph
e
no
m
ena
at
tribu
te
d
to
highe
r
swi
tc
hin
g
f
reque
nc
ie
s.
Fiel
d
distrib
ution
at
dif
f
eren
t
a
ir
-
ga
p
le
ng
t
hs
betwee
n
twist
s
with
e
nam
el
t
hick
ness
of
40
µ
m
at
diff
ere
nt
switc
hing
f
re
qu
e
ncies
in
the
range
of
10
-
30
kHz
is
nu
m
erical
ly
m
od
el
ed
usi
ng
COMS
OL.
The
sim
ulati
on
s
car
ried
out
sho
wed
th
at
the
el
ect
ric
fie
l
d
distrib
ution d
e
pends
on t
he
ai
r
-
gap le
ngth
be
tween t
he wire
s and i
ns
ulati
on thic
kn
e
ss.
The
rest
of
th
e
pap
e
r
is
organ
iz
e
d
as
f
ollow
s:
Sect
io
n
2
descr
i
bes
the
ex
per
im
ental
set
up
f
or
br
ea
kdow
n
te
s
ts,
sam
ple
pr
e
par
at
io
n
,
an
d
te
st
proce
dure.
Sect
ion
3
disc
us
ses
the
e
xpe
rim
ental
resu
lt
s
an
d
analy
ze
d
the
infl
uen
ce
of
hig
h
s
witc
hing
f
reque
ncy
on
t
wiste
d
pai
r
sa
m
ples
by
m
od
el
ing
the
el
ect
ric
fiel
d
distrib
ution
wi
th
ap
plied
vo
l
ta
ge
an
d
re
du
ced
brea
kdow
n
volt
age.
Fin
al
ly
,
sect
ion
4
con
cl
ud
e
s
wi
th
the
fin
dings
on
br
eakdo
wn
volt
age
a
nd
rate
of
deterio
rati
o
n
of
t
he
ins
ulati
on
syst
em
fo
r
Square
wa
vefor
m
,
Square
-
risin
g
wav
e
f
or
m
,
and
Squa
re
-
s
pik
e
wav
e
f
or
m
s.
2.
E
X
PERI
MEN
TAL PR
O
CEDURE
2.1. S
amp
le
p
repar
at
i
on
The
t
wiste
d
-
pa
ir
sp
eci
m
ens
a
re
m
ade
of
tw
o
e
nam
e
l
ed
m
agn
et
wires
tw
ist
ed
li
ke
a
pla
it
as
pe
r
th
e
AS
TM
D
1676
-
03
sta
ndar
ds
.
A
sin
gly
coated
poly
est
er
ena
m
el
insu
la
ti
on
with
40
µm
thickne
ss
is
e
m
plo
ye
d
for
te
sti
ng
.
T
he
twist
ed
-
pair
le
ng
th
is
m
ai
ntained
(
12±
6)
c
m
as
per
sta
ndar
ds
.
T
he
num
ber
of
twist
s
is
6,
dep
e
ndin
g on the
diam
et
er o
f t
he
wi
re
.
2.2
.
Te
st
s
et
up
Figure
1
sho
w
s
the
ex
per
im
e
ntal
set
up
for
Square
,
Squa
re
-
risin
g,
a
nd
Square
-
s
pik
e
waveform
s
with
high
-
fr
e
quency
,
the
high
volt
age
us
e
d
f
or
te
sti
ng
the
twist
ed
pair
sam
ples.
All
the
waveform
s
us
ed
he
re
are
bipolar
in
natu
re.
The
23
0V,
50Hz
source
is
fed
to
the
EMI
li
ne
filt
er
and
the
out
pu
t
is
con
nected
t
o
the
fr
e
qu
e
ncy
an
d
du
ty
cy
cl
e
cont
ro
ll
er.
T
he
outpu
t
of
the
inv
e
rter
is
ste
pp
e
d
up
by
e
m
plo
yi
ng
a
na
nocryst
al
li
ne
cor
e
tra
ns
f
orm
er
for
high
fr
e
quency
operati
on
w
here
the
t
wiste
d
pair
sam
ple
is
stresse
d.
A
dig
it
al
os
ci
ll
os
co
pe
m
easur
e
s
the
pe
ak
vo
lt
age
th
r
ough
a
1000:1
capaci
ta
nce
volt
age
div
ide
r
.
The
la
borato
r
y
set
up
us
e
d
f
or c
ondu
ct
ing
the
brea
kdow
n
a
nd w
it
hst
and test
s is s
how
n
in
Fig
ure
2
.
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Mo
deling of
hi
gh freq
ue
ncy
hi
gh
v
olta
ge of
wavef
orms
on l
if
e o
f en
am
el
insu
l
atio
n
(
S
N
ar
asi
m
ha R
ao
)
1333
Figure
1.
Ex
pe
rim
ental
setup
Figure
2
.
Lab
orat
or
y
set
up
use
d for
br
ea
kdown
te
sts
2.3
.
Te
s
t
p
r
oc
edure
Break
dow
n
te
s
ts
wer
e
ca
rr
ie
d
out
on
the
tw
ist
ed
-
pai
r
s
pec
i
m
ens
with
the
ed
ge
of
on
e
wire
is
fe
d
with
te
st
volt
age
a
nd
e
dge
a
nothe
r
wi
re
is
gro
unde
d
th
us
t
he
la
ye
rs
of
t
he
enam
el
of
bo
th
wire
s
com
es
unde
r
high
volt
age
stress.
T
he
br
ea
kdown
te
st
is
perform
ed
by
t
he
ap
plica
ti
on
of
a
high
volt
age
at
the
twisted
-
pa
i
r
sp
eci
m
en
un
ti
l
the
sam
ple
in
su
la
ti
on
brea
kdow
ns
.
The
wi
thstan
d
te
st
wa
s
co
nducte
d
to
est
i
m
at
e
the
ti
m
e
up
to w
hic
h
the in
su
la
ti
on
withstands wit
hout break
dow
n
eve
n under
the a
pp
l
ic
at
ion
o
f vari
ous volt
age leve
ls. In
this
stu
dy,
t
he
br
ea
kdow
n
vo
l
ta
ge
of
t
he
t
wiste
d
-
pair
sp
eci
m
en
is
obse
rv
e
d
by
e
nh
a
ncin
g
th
e
vo
lt
age
a
t
a
rat
e
of
500
V
olt/
sec
up
to
br
e
ak
dow
n.
The
brea
kdown
ex
per
i
m
ents
wer
e
do
ne
on
i
nsula
ti
on
with
enam
el
at
room
tem
per
at
ur
e
unde
r
the
a
ppli
cat
ion
of
a
hi
gh
-
f
re
qu
e
ncy
high
volt
age
of
S
qu
a
re,
S
qua
re
-
rising,
an
d
Square
-
sp
ike
wa
veform
s
and
a
fr
e
qu
ency
in
the
ra
nge
of
10
-
30
kHz.
Br
eak
dow
n
volt
age
is
ta
ken
as
the
a
ve
rag
e
of
five
br
ea
kdow
n
val
ues.
It
is
no
ti
ced
t
hat,
w
hile
perform
ing
the
e
xp
e
rim
e
nts
at
high
-
fr
e
qu
e
ncy,
the
i
oniz
at
ion
i
m
pact
betwee
n
t
he
twist
e
d
-
pa
ir
ga
p
s
is
m
a
nifested
in
the
f
or
m
of
a
haz
y
vio
le
t
glow.
Re
du
ce
d
break
dow
n
vo
lt
age
is
a
pp
li
ed
to
the
sa
m
ple
un
ti
l
the
br
ea
kdown
of
the
sam
ple
fo
r
th
e
W
it
hs
ta
nd
-
te
st.
F
or
a
certai
n
fr
e
qu
e
ncy,
the
red
uce
d
volt
a
ge
ap
plied
is
75
%,
60%
,
an
d
50
%
of
the
break
dow
n
volt
a
ge
of
that
pa
r
ti
cular
fr
e
qu
e
ncy.
Tab
le
1
s
hows
the
par
am
et
ers
of t
he
te
st set
up
use
d for
br
ea
kdown
test
s.
Table
1.
T
est
s
et
up
detai
ls
Frequ
en
cy
Du
ty
c
y
cle
W
av
ef
o
r
m
s
Test op
erating
vo
ltag
e
10
-
3
0
kHz
0
-
50%
Sin
e,
Sq
u
are, Sq
u
are
-
risin
g
,
an
d
Squ
are
-
sp
ik
e
0
-
1
5
kV p
eak
3.
RESU
LT
S
AND DI
SCUS
S
ION
3.1. Ex
peri
ment
al
r
esults
The
var
ia
ti
on
i
n
br
ea
kdow
n
vo
lt
age
s
ov
e
r
t
he
f
re
qu
e
ncy
with
wav
e
f
orm
s
is
dep
ic
te
d
in
Fi
gure
3.
Fr
om
the
gr
a
ph,
it
is
eviden
t
that
the
Square
,
Square
-
risi
ng
and
S
quare
-
spi
ke
wa
vefor
m
s
exh
i
bit
two
strai
ght
li
nes
with
va
r
yi
ng
sl
op
es
as
in
[
19]
,
with
a
rise
in
the
frequ
e
ncy
from
10
-
30
kHz.
It
is
obser
ve
d
f
r
om
the
gr
a
phs
t
he
br
e
akdo
wn
vo
lt
ag
e
of
Square
-
ri
sing
wa
ve
for
m
is
m
or
e
than
that
of
Squ
are
-
s
pik
e
an
d
Square
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
3
,
Se
ptem
ber
20
21
:
13
3
1
-
133
9
1334
wav
e
f
or
m
s.
Ther
ef
or
e
,
it
is
de
scribe
d
that
the
deterior
at
i
on
rate
of
the
ins
ul
at
ion
syst
e
m
o
f
twist
ed
pair
usi
ng
Square
wa
vefor
m
is
qu
ic
ke
r
as
c
om
par
ed
to
S
quare
-
s
pik
e
a
nd
S
qu
a
r
e
-
risin
g
wa
vefor
m
s,
sim
il
ar
to
th
e
publishe
d
res
ul
ts
fo
r
P
WM
li
ke
an
d
P
W
M
peak
wavef
orm
s
rep
ort
e
d
in
[20].
An
at
te
m
pt
is
m
ade
to
com
par
e
the
res
ults
ob
t
ai
ned
f
or
Squa
re
sp
i
ke
with
t
hat
of
publishe
d
res
ults
of
P
WM
pea
k
wa
ve
form
.
It
cou
ld
be
that
the
S
quare
wa
ve
has
a
ste
e
p
rising
a
nd
trai
li
ng
ed
ge
t
hat
c
orres
ponds
to
high
fr
e
quency
com
ponen
ts
due
t
o
t
he
sh
a
pe
of
t
he
wav
e
f
or
m
.
Ther
e
a
re
quit
e
diff
e
re
nt
pola
ri
zi
ng
a
nd
relaxi
ng
e
ff
ect
s
on
di
po
le
s
f
or
a
co
m
ple
te
Square
wa
ve
per
i
od
[21].
It
is
becau
s
e
of
this
reas
on
t
hat
there
will
be
no
possib
il
ity
of
m
or
e
dipoles
fo
ll
owin
g
the
r
apid
re
ve
rsal
of
the
pola
ri
ty
.
I
t
is
al
so
ob
se
rved
that
re
du
ct
i
on
i
n
break
do
wn
volt
age
is
27%
in
the
S
qu
a
re
wa
vefor
m
wh
e
rea
s
32%
i
n
S
qu
a
re
-
s
pik
e
wa
vefor
m
and
41%
i
n
S
quare
-
risin
g
wav
e
form
wh
en
th
e
fr
e
qu
e
ncy
is
va
ried
f
ro
m
10
-
30
kH
z
.
Fig
ure
4
sh
ows
the
aging
te
st
res
ults
stressed
wit
h
a
Square
-
risin
g
wav
e
f
or
m
with
var
io
us
le
vels
in
fr
e
quency
up
to
30
kH
z
at
reduce
d
volt
ag
es.
It
can
be
se
en
that
the
dec
rease
in
brea
kdown
t
i
m
e
fo
ll
ow
s
a
higher
slo
pe
w
hen
the
a
ppli
ed
te
st
volt
age
is
cha
nged
fro
m
75
%
to
60%
an
d
a
lowe
r
slo
pe
w
hen
cha
nged
from
60
%
to
50
%
for
al
l
thr
ee
fr
e
qu
e
ncies
th
ough
the
c
ha
nge
in
sl
ope
is
ve
ry
le
ss
at
10
kHz.
It
i
s
interest
in
g
to
note
that
re
sul
ts
in
Fig
ur
e
4(
a
)
for
S
quare
-
sp
ike
wa
veform
po
int
to
a
si
m
il
ar
tren
d
to
that
of
publishe
d
res
ul
ts
fo
r
P
WM
-
pe
ak
wa
vefo
rm
repor
te
d
in
[
12
]
,
fo
r
al
l
three
fr
e
qu
e
ncies,
t
houg
h
the r
es
ults
of
F
igure
4(
b
)
is in
contrast.
This
m
ay
b
e due t
o hig
her te
st v
oltages i
n
S
quare
-
risin
g
wa
veform
s
.
Figure
3
.
V
a
riat
ion
in
brea
kdown
volt
ages
of
s
quare,
s
qu
a
re
-
rising
, a
nd
s
quare
-
s
pik
e
wa
v
e
form
s
fr
om
10
-
30
kHz
(a)
(b)
Figure
4
.
A
ging test
r
e
su
lt
s at
r
e
du
ce
d v
oltage
up to 3
0 kHz
stresse
d wit
h
(a)
square
-
s
pike
w
a
vefor
m
(
b)
sq
ua
re
-
risin
g
wav
e
f
or
m
Figure
5
(a)
-
(c
)
sh
ows
the
agi
ng
te
st
resu
lt
s
wh
e
n
the
twist
ed
pair
sam
ple
is
stressed
with
Square
-
rising
a
nd
S
qu
are
-
s
pik
e
wav
e
form
s
with
fr
equ
e
ncy
le
vels
of
10
kH
z
,
20
kH
z
,
an
d
30
kHz
res
pecti
vel
y
with
a
reduce
d
vo
lt
ag
e
of
75%,
60
%,
a
nd
50%
.
Wh
e
n
t
her
e
is
an
en
ha
ncem
ent
in
f
reque
nc
y
from
10
-
30
kHz,
Square
-
risin
g
and
Square
-
s
pi
ke
wa
vefo
rm
s
sh
ow
the
t
wo
strai
gh
t
-
li
ne
slop
e
s,
w
he
reas
auth
or
s
of
[
13
]
,
[22
]
-
[
24]
,
repor
te
d
un
if
or
m
slope
s
of
c
har
act
e
risti
cs
with
va
rio
us
e
nam
el
s
f
or
fr
e
quenci
es
up
t
o
8
kHz.
The
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Mo
deling of
hi
gh freq
ue
ncy
hi
gh
v
olta
ge of
wavef
orms
on l
if
e o
f en
am
el
insu
l
atio
n
(
S
N
ar
asi
m
ha R
ao
)
1335
char
act
e
risti
cs
of
the
Square
-
sp
ike
wa
vefo
r
m
are
ob
se
rv
e
d
to
cr
os
s
ove
r
the
cha
racter
ist
ic
s
of
t
he
S
qu
a
re
-
rising
wa
veform
on
ly
for
10
kH
z
i
n
Fig
ure
5(
a
).
T
he
br
ea
kdown
occ
ur
s
early
for
the
S
qu
a
re
-
risin
g
w
avefor
m
than
t
he
S
qu
a
r
e
-
sp
i
ke
wav
e
form
du
e
to
the
hi
gh
e
r
ap
plied volt
age
in
al
l
th
e
fr
e
quencies.
Th
e reason f
or low
e
r
br
ea
kdow
n
volt
age
is
that
at
higher
s
witc
hi
ng
fr
e
quency
t
he
io
ns
getti
ng
trapp
e
d
betwe
en
the
el
ect
r
odes
and
rap
i
d
pola
rity
changes
,
res
ulti
ng
in
the
dist
ort
ed
el
ect
ro
s
ta
t
ic
fiel
d.
Du
e
to
this
sp
ace
cha
rg
e
no
lo
nger
f
ollow
the
fiel
d
du
e
t
o
the
eff
ect
of
pola
rizat
ion.
P
D
m
ay
be
influ
e
nced
by
surfac
e
chargin
g
ca
use
d
pre
-
i
oniz
at
ion
of
el
ect
ric fiel
d
st
rength
.
(a)
(b)
(c)
Figure
5
.
A
ging test
c
har
act
er
ist
ic
s o
f
s
quare
-
risin
g
a
nd
s
quare
-
s
pik
e
wa
ve
form
s at fr
eq
ue
ncies
of
(
a)
10 kHz
,
(b
) 20 k
Hz, an
d
(
c)
30
kH
z
Figure
6
sho
w
s
the
brea
kdown
ti
m
e
wh
en
stressed
with
a
reduce
d
vo
lt
age
of
Squa
r
e
-
risin
g
a
nd
Square
-
s
pik
e
wav
e
f
or
m
s
up
to
30
kH
z
.
Fro
m
Figu
re
6
(a
)
it
is
seen
th
at
br
ea
kdown
ti
m
e
do
esn
’t
show
a
ny
change
w
he
n
the
fr
e
quency
is
enh
a
nced
fro
m
10
-
30
kHz
wh
e
n
stresse
d
with
Square
-
ri
sing
wa
ve
form
at
a
reduce
d
volt
a
ge
of
75%
(
4.2
3
kV).
H
owever,
un
der
t
he
reduce
d
vo
lt
age
stres
s
of
60%
(
3.3
9
kV
)
,
the
br
ea
kdow
n
tim
e
is
decr
ea
sed
by
64%
wh
e
n
t
he
f
re
qu
e
ncy
is
en
han
ce
d
fro
m
10
-
20
kHz
a
nd
by
24
%
f
rom
20
-
30
kHz
.
Sim
i
l
arly
,
at
50%
(
2.8
2
kV)
br
ea
kdow
n
ti
m
e
is
decr
eased
by
52
%
w
he
n
the
f
r
equ
e
ncy
is
e
nh
anced
from
10
-
20
kHz
an
d
de
c
rea
sed
by
47%
w
hen
the
fr
e
qu
e
ncy
is
enh
a
nce
d
from
20
-
30
kH
z
.
This
sho
ws
the
eff
ect
of
f
re
quency
at
10
kH
z
an
d
20
kH
z
com
par
ed
t
o
30
kH
z
at
reduc
ed
vo
lt
a
ges.
It
is
seen
f
ro
m
Figure
6
(b)
that
br
ea
kd
own
tim
e d
ecreased w
it
h
a
hi
gh
e
r
slo
pe
w
he
n
the f
r
eq
ue
nc
y i
s
enh
ance
d
f
ro
m
1
0
-
20
kH
z
and
a
lowe
r
slo
pe
f
ro
m
20
-
30
kHz
wh
e
n
t
he
reduce
d
volt
age
of
75%
(
3.3kV)
f
or
S
quare
-
s
pik
e
wa
vefor
m
.
Howe
ver,
w
he
n
the
volt
age
of
60%
(
2.64
kV),
the
break
dow
n
ti
m
e
is
decr
eased
by
31
%
w
hen
the
frequ
e
ncy
is
en
han
ce
d
f
r
om
10
-
30
kH
z
.
Sim
i
la
rly
,
at
50%
(
2.2
kV)
br
ea
kdow
n
ti
m
e
is
decr
ease
d
by
30%
w
he
n
t
he
fr
e
qu
e
ncy
is
e
nh
a
nce
d
from
10
-
30
kHz.
T
he
si
gn
ific
a
nc
e
of
the
fin
di
ng
is
that
t
he
twist
ed
pair
enam
e
l
insu
la
ti
on u
nd
ergoes sev
e
re
e
le
ct
rical
stress at
high f
re
quen
ci
es.
At
10 k
H
z,
the
break
do
wn
stre
ng
t
h
is 1.15
t
o
4.4
kV
ti
m
es
hig
he
r,
1.12
to
3.4
kV
ti
m
es
hig
he
r
at
20
kH
z,
an
d
at
30
kHz
1.25
to
2.1
kV
ti
m
es
hig
he
r
f
or
Square
,
Squar
e
-
risin
g
,
a
nd
Square
-
s
pik
e
wav
e
f
or
m
s
com
par
ed
to
resu
lt
s
repor
te
d
in
[12
]
,
[
2
3],
in
the
li
te
ra
ture
f
r
om
2.4 kHz
to
8
kHz.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
3
,
Se
ptem
ber
20
21
:
13
3
1
-
133
9
1336
(a)
(b)
Figure
6
.
The
c
hange i
n
ti
m
e
with the
ch
a
ng
e in
diff
e
re
nt fr
equ
e
ncy le
ve
ls u
p 30 kHz
(a)
sq
ua
re
-
risin
g
wav
e
f
or
m
b
)
s
qu
a
re
-
sp
i
ke wa
vefor
m
3.2.
Nu
meri
c
al m
od
el
ing
of
e
le
ctric fiel
d
distribu
tion i
n
twi
sted pair
s
amp
le
In
ge
om
et
rical
l
y
si
m
plifie
d,
a
par
al
le
l
c
onfig
ur
at
io
n
of
windin
g
wires
wit
hout
t
wists,
th
r
ee
cases
a
re
analy
zed; case
1
:
Two
c
onduc
tors
are in
dire
ct
co
ntact
w
it
h
each o
ther
(
g=
0
m
m
),
case
2
:
Two
con
du
ct
ors are
separ
at
e
d
by
a
m
ic
ro
ga
p
be
tween
them
(g=0.
05
m
m
)
and
case
3
:
Tw
o
cond
ucto
rs
are
separ
at
e
d
by
a
ga
p
(g
=
0.1m
m
).
The
el
ect
ric
fiel
ds
of
tw
o
neighb
or
i
ng
wire
s
wer
e
num
eri
cal
ly
m
od
el
ed
us
in
g
the
C
O
MSOL
pro
gr
am
.
Fo
r
m
od
el
ing
,
the
po
ly
est
er
ena
m
el
thickness
was
co
ns
ide
re
d
to
be
40
µm
,
and
the
diele
c
tric
const
ant
was
a
ssu
m
ed
as
€
r
=3.3.
The
ap
pl
ie
d
volt
age
is
the
brea
kdown
volt
age
of
th
e
Squa
re
-
risin
g
a
nd
Square
wa
vefo
rm
s
at
var
iou
s
le
vels
in
f
re
quency
in
the
ra
nge
of
10
-
30
kHz
obse
r
ved
i
n
the
ex
pe
rim
en
t
.
The
el
ect
ric
fiel
d
di
stribu
ti
on
of
th
e
2D
m
od
el
of
two
par
al
le
l
c
onduct
ors
for
al
l
the
3
cases
(
g=
0
m
m
,
g=0
.
05
mm
&
g=
0.1
m
m
)
is
stresse
d
by
the
S
quare
-
r
isi
ng
wa
veform
a
t
a
fr
e
que
ncy
of
10
kHz
.
T
he
el
ect
ri
c
fiel
d
distrib
ution
of
the
2D
m
od
el
stressed
by
the
Square
wav
e
f
or
m
at
a
fr
eq
ue
ncy
of
10
kH
z
is
si
m
i
la
r
to
the
2D
m
od
el
design
e
d
with
a
Squa
re
-
rising
wa
ve
form
.
Figu
re
7
s
hows
the
e
le
ct
ric
fiel
d
di
stribu
ti
on
of
t
he
2D
m
od
el
f
or case
1 (g =0
m
m
)
of
the Sq
uar
e
-
risi
ng w
a
ve
form
.
Figure
7
.
The
e
le
ct
ric f
ie
ld
distr
ibu
ti
on
of s
quare
-
risin
g wa
vefor
m
at 10 kHz
for
case
1
I
t
is
no
te
d
that
for
al
l
the
3
ca
ses,
the
m
axi
m
um
streng
th
of
the
el
ect
ric
fiel
d
is
on
t
he
sur
face
of
the
wire
ins
ulati
on
near
the
co
ntact
locat
ion
of
the
wires
.
In
ca
se
2,
t
he
reas
on
for
r
e
du
ce
d
el
ect
ric
fiel
d
st
rength
is
because
of
t
he
narrow
ai
r
gap
with
t
hat
of
the
direct
c
onta
ct
case
.
In
c
ase
3,
it
is
note
d
that
t
he
rea
so
n
f
or
m
or
e
red
ucti
on
of
fiel
d
stre
ng
t
h
is
due
to
the
ai
r
ga
p
be
tween
th
e
wi
res
com
par
e
d
to
case
1
a
nd
case
2.
Howe
ver
,
the
e
le
ct
ric
fiel
d
on
the
s
urface
of
the
wi
re
in
su
la
ti
on
betwee
n
t
he
wires
is
sti
ll
suffici
ent
t
o
i
niti
at
e
PD
f
or
both
th
e
wa
vefor
m
s
be
cause
i
on
iz
at
i
on
is
s
urrou
nd
i
ng
the
ai
r
ga
p
al
ong
with
s
ur
face
po
te
ntial
decay
[
25]
.
A
sim
il
ar
trend
of
res
ults
repor
te
d
in
[
18]
.
For
the
incr
eased
volt
age
of
5.6
5
kV,
the
fiel
d
value
obta
ine
d
is
46
.
6
kV/m
m.
If
instea
d
of
1kV
te
st
volt
age
,
5.65
k
V
is
ap
plied
the
fiel
d
would
hav
e b
e
en
39.
55
k
V/m
m
fo
r
the
sam
e
0.
1
m
m
gap
.
How
ever,
the
fiel
d
ob
ta
ine
d
by
[
18]
,
is
7
kV
/m
m
fo
r
t
he
te
st
vo
lt
age
of
1kV
.
Hen
ce
,
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Mo
deling of
hi
gh freq
ue
ncy
hi
gh
v
olta
ge of
wavef
orms
on l
if
e o
f en
am
el
insu
l
atio
n
(
S
N
ar
asi
m
ha R
ao
)
1337
the
hi
gh
e
r
fiel
d
ob
ta
ine
d
i
n
t
his
w
ork
is
du
e
to
the
hi
gh
f
reque
ncy
of
10
kHz
c
om
par
ed
to
po
wer
fr
e
qu
e
nc
y
and
S
qu
a
re
-
rising
wa
ve
c
om
par
ed
t
o
Si
ne
w
ave.
Be
si
des
th
is,
the
€
r
val
ue
of
t
he
m
at
erial
us
e
d
by
[
18
]
,
i
s
3.8
-
4.2
com
pa
red
to
the
poly
est
er
m
a
te
rial
us
ed
in
this
w
ork
fo
r
w
hich
the
€
r
value
us
e
d
is
3.
3.
T
he
sa
m
e
is
ob
s
er
ved
f
or
20
kH
z
an
d
30
kH
z
.
T
his
s
ho
ws
t
he
ef
fect
of
high
f
re
qu
e
nc
y
and
w
ave
form
s
on
th
e
distrib
utio
n
of
fiel
d
in
the
ai
r
gap
betwee
n
the
wires
an
d
al
so
on
it
s
s
urface.
T
his
ca
n
io
nize
the
ai
r
in
the
gap
ca
us
i
ng
sp
ar
k
le
adin
g
to
a
therm
al
ru
naw
ay
as
well
.
So
besid
es,
th
e
PD
init
ia
te
d
by
high
fr
e
que
ncy
stress,
io
nizat
ion
m
ay
le
ad
to
s
park
i
n
the
ga
p
th
us
trig
ger
i
ng
the
break
dow
n
of
e
nam
e
l
insu
la
ti
on.
T
her
e
is
a
need
f
or
t
he
dev
el
op
m
ent of a
ne
w
ins
ulati
on syst
em
f
or
l
ow volt
age m
oto
rs
t
hat can
w
i
thstan
d hig
her
PD
resist
ance
.
Figure
8
s
how
s
that
there
is
a
var
ia
ti
on
in
the
distri
bu
ti
on
of
el
ect
ric
fi
el
d
over
t
he
f
r
equ
e
ncy
of
Square
-
risin
g
a
nd
S
quare
waveform
s
in
all
t
hr
ee
cases
.
For
cases
1,
2
an
d
3
the
el
ect
ric
fiel
d
streng
t
h
is
hig
h
at
10
kHz,
a
nd
a
grad
ual
de
crease
in
el
ect
ric
fiel
d
stre
ngth
values
at
20
kH
z
a
nd
30
kHz
f
or
bo
t
h
S
qu
a
re
-
rising
a
nd
S
qu
are
wavef
or
m
s.
Fu
rt
her
,
it
is
no
ti
ced
that
f
or
al
l
three
cases,
the
el
ect
ric
fiel
d
stren
gth
of
the
Square
-
risin
g
wav
e
f
or
m
is h
igh
e
r
t
han that
of
the
Square
wav
e
f
or
m
for
al
l fr
eq
ue
ncies
.
Figure
8
.
Va
riat
ion
in
elec
tric
fiel
d dist
rib
ution
of S
quare
-
ri
sing
an
d
s
quar
e w
a
vefor
m
s f
r
om
1
0 kH
z
to 30
k
Hz
f
or case
1, case
2,
a
nd ca
se 3
3.3.
Nu
meri
c
al m
od
el
ing
of
e
le
ctric fiel
d
distribu
tion
of
reduced
volt
ages in
twi
sted
pa
ir
s
amp
le
A
n
um
erical
m
od
el
fo
r
re
duced
vo
lt
a
ge
of
75%,
60%
,
and
50%
of
t
he
br
ea
kdown
volt
age
of
the
Square
-
risin
g
wav
e
f
or
m
is
con
si
der
e
d.
T
he
va
riat
ion
i
n
t
he
el
ect
ric
fiel
d
distrib
ution
ov
e
r
t
he
f
re
quency
of
reduc
e
d
volt
ag
es
is
sh
own
in
Figure
9
.
How
ever,
the
el
ect
ric
fiel
d
ob
se
rved
is
hig
h
e
nough
to
init
ia
te
PD
f
or
75%,
60%, a
nd
50%
of r
e
duce
d vo
lt
age
.
Figure
9
.
Va
riat
ion
in
elec
tric
fiel
d dist
rib
ution
of S
quare
-
ri
sing wa
vefo
rm
unde
r red
uce
d v
oltage
from
1
0
kH
z
to 3
0
kH
z
for
case
1, case,
and case
3
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
3
,
Se
ptem
ber
20
21
:
13
3
1
-
133
9
1338
4.
CONCL
US
I
O
N
The
high
s
witc
hing
fr
e
qu
e
ncy
is
a
pr
e
do
m
inant
fact
or
that
causes
reli
abili
ty
prob
le
m
s
in
ins
ulati
on
and
ca
n
be
ass
ociat
ed
with
P
D
act
ivit
y
that
acce
le
rates
insu
la
ti
on
deteri
or
at
io
n
of
the
m
oto
r
wind
i
ng.
Af
te
r
inv
est
igati
ng
the
ex
per
im
ental
resu
lt
s
it
is
rev
eal
ed
that
f
or
Square
-
risi
ng
wav
ef
orm
the
br
ea
kdow
n
vo
l
ta
ge
is
high
w
hen
c
om
par
ed
to
Sq
ua
re
-
s
pik
e
a
nd
Square
w
a
ve
form
s.
It
is
al
so
ob
se
rv
e
d
that
wh
il
e
us
in
g
S
qu
a
re
wav
e
f
or
m
the
rate
of
dete
riorat
ion
of
the
ins
ulati
on
syst
em
is
exp
e
diti
ou
s
wh
e
n
com
par
e
d
with
S
qu
a
re
-
sp
ik
e
and
S
quare
-
ris
ing
wa
ve
form
s.
Furtherm
ore,
it
is
a
lso
no
ti
ced
that
wh
e
n
the
te
st
vo
lt
age
f
requ
en
cy
is
enh
a
nce
d
f
ro
m
10
-
30
kHz,
th
e
li
fetim
e
of
Sq
ua
re
-
risin
g
an
d
S
qu
a
re
-
sp
i
ke
wav
e
f
or
m
s
are
two
strai
gh
t
li
nes
with
hi
gh
e
r
sl
op
e
f
ollow
e
d
by
lowe
r
slo
pe
for
al
l
fr
e
que
ncies.
F
ro
m
the
si
m
ulati
on
,
it
is
ob
se
rv
e
d
t
hat
the
el
ect
ric
fiel
d
va
lues
are
highe
st
in
the
ga
p
be
tween
the
wire
s
includi
ng
at
the
surface
of
wires,
a
nd
dec
r
easi
ng
on
th
e
oth
e
r
re
gions.
T
his
im
plies
that
the
possibil
it
y
of
P
D
init
ia
ti
ng
wi
thin
the
enam
el
.
Fr
om
the
si
m
ula
ti
on
resu
lt
s,
it
is
a
na
ly
zed
that
the
i
m
pact
of
PD
on
enam
el
insu
la
ti
on
a
t
high
er
s
witc
hing
f
r
equ
e
ncies
is
s
o
high
that
the
twist
ed
-
pair
sam
ple
fail
s
pr
em
at
ur
el
y.
Hen
ce,
t
here
is
a
need
f
or
further
i
nv
e
sti
gation
of
desi
gnin
g
m
agn
et
ic
wire
s
to
withstan
d
high
PD
resist
ance
in
lo
w
volt
age
m
oto
rs.
The
i
m
po
rta
nt
sta
ti
st
ic
al
par
a
m
et
e
r
su
c
h
as
PDIV
, PR
PD
patte
r
n st
ressed by
Square
w
a
vefo
rm
n
ee
ds
t
o be in
vestigat
ed
.
REFERE
NCE
S
[1]
V.
N.
Höpner,
V.
E.
W
ihelm
,
“
Insulat
ion
Li
f
e
Span
of
Low
Volta
ge
El
e
ct
r
ic
Motors
-
A
Surv
e
y
,
”
Journal
o
f
ene
rgies,
vol
.
14
,
pp
.
1
-
32,
2021
,
doi:
10
.
3390/e
n1
4061738.
[2]
F.
Guastavi
no,
C.
Gianogl
io
,
E.
Tore
ll
o
,
M.
Ferra
ris
,
and
W
.
Giane
l
li
,
“
Beha
v
io
ur
of
Convent
io
nal
and
Nanofi
l
l
ed
Im
pre
gnat
ion
R
esins
when
Subjected
to
PD
Acti
vi
t
y
,”
IEEE
2
nd
Inte
rnational
Confe
renc
e
on
Diel
e
ct
rics
IC
D
,
2018,
pp
.
1
-
3
,
d
oi:
10
.
1109/ICD.
2018.
8514692.
[3]
L.
Lusuar
di
,
A.
Cava
llini
,
M.
G.
de
la
Calle,
J.
M.
Martí
ne
z
-
Tarifa
,
G.
Robl
es,
“
Insulat
ion
desi
gn
of
low
volt
age
el
e
ct
ri
ca
l
m
otors
fed
b
y
PW
M
in
ver
te
rs,
”
IEEE
E
le
c
tric
al
Insulation
Magazine
,
vol.
35,
no.
3,
pp.
7
-
15,
Ma
y
-
Jun
e
2019,
doi
:
10
.
11
09/MEI.
2019.
86
89431.
[4]
Y.
Xie,
J.
Zha
ng
,
F.
L
eona
rdi
,
A.
R.
Munoz,
M.
W
.
Degne
r,
F.
L
ia
ng,
“
Volt
age
S
tre
ss
Modeli
ng
a
nd
Mea
surem
ent
for
Random
-
W
ound
W
indi
ngs
Drive
n
b
y
Inv
er
te
rs,”
IE
EE
In
ter
nati
onal
E
lectri
c
Mac
h
ine
s
&
Dr
iv
es
Confe
ren
ce
IE
MDC
,
2019
,
p
p.
1917
-
1924
,
d
oi:
10
.
1109/IE
M
DC.2019.
87851
33.
[5]
R.
Le
u
zz
i
,
V.
G.
Monopoli,
F.
Cupert
ino
,
P.
Z
anc
he
tt
a
,
“
Acti
v
e
Ageing
Contr
ol
of
W
indi
ng
Insulat
ion
in
Hig
h
Freque
nc
y
E
lec
tri
c
Driv
es,
”
I
EE
E
Ene
rgy
Conve
rs
ion
Co
ngress
and
Expos
it
ion
ECCE
,
2018,
pp.
1
-
7,
doi:
10
.
1109/E
C
CE.
2018.
855825
7.
[6]
Z.
W
ei,
H.
You,
B.
Hu,
R.
N
a,
J
.
W
ang,
“
Partial
Discha
rge
B
ehavior
on
Twiste
d
Pair
under
Ultr
a
-
short
Rise
Ti
m
e
Square
-
wave
Exc
itat
ions,”
I
EE
E
E
lectric
a
l
Insulati
on
Confe
renc
e
EI
C
,
2019,
pp.
493
-
496,
doi
:
1
0.
1109/E
IC432
17.
2019.
904662
4.
[7]
G.
C.
Montana
ri
,
F.
Negri
,
F.
Ci
ani
,
“
Parti
a
l
Discha
rge
and
l
ife
beha
vior
of
rotat
ing
m
ac
hine
wir
e
insula
ti
on
und
er
PW
M
wave
forms:
The
infl
u
ence
of
inv
erter
ch
ar
ac
t
eri
sti
cs,
”
IEEE
E
le
c
tric
al
Ins
ulat
ion
Con
fe
re
nce
EIC
,
2017,
p
p.
161
-
164,
doi
:
10
.
1109/E
IC
.
2017.
8004619.
[8]
G.
C.
Montana
r
i,
P.
Seri,
“
The
eff
ec
t
of
inve
r
t
er
cha
ra
cteri
sti
c
s
on
par
ti
al
discha
rge
and
l
ife
beha
vior
of
wire
insula
ti
on
,
”
IE
EE
El
e
ct
ri
cal
Insulati
on
Magazine
,
vol.
34,
no.
2,
pp.
32
-
39,
Marc
h
-
April
2018,
doi
:
10.
1109/MEI.
20
18.
8300442.
[9]
M.
G.
De
La
Cal
le,
J.
M.
M
art
ín
ez
-
T
ari
f
a,
Á.
M.
Góm
ez
Solani
lla,
G.
R
oble
s,
“
Unce
rt
aint
y
Source
s
in
the
Esti
m
at
ion
of
th
e
Parti
a
l
Discha
r
ge
Inc
eption
Volta
ge
in
Turn
-
to
-
Turn
Insulat
ion
S
y
stems
,
”
IEEE
Ac
c
ess
,
vol.
8,
p
p.
157510
-
157519,
2020,
doi: 10.
11
09/ACCESS
.
2020.
3018870.
[10]
R.
Le
u
zz
i
,
V.
G.
Monopoli,
L
.
Rovere
,
F.
Cup
ert
ino
,
P.
Za
n
ch
et
t
a,
“
Anal
y
s
is
and
Detect
ion
o
f
El
ectri
ca
l
Agi
ng
Eff
ects
on
High
-
Speed
Motor
Insulat
ion
,
”
IEEE
Tr
ansacti
ons
on
Industry
Appl
ic
at
ions
,
vol.
5
5,
no.
6,
pp.
60
18
-
6025,
Nov.
-
De
c. 2019, doi
:
10.
11
09/T
IA.2019.
29
29013.
[11]
K.
Ham
e
y
er,
A.
Ruf,
F.
Paul
i,
“
Influe
nce
of
fa
st
sw
it
chi
ng
se
m
ic
onduct
ors
o
n
the
winding
i
nsulat
ion
s
y
s
tem
of
el
e
ct
ri
ca
l
m
ac
hi
nes,
”
In
te
rnat
ion
al
Pow
er
Elec
tr
onic
s
Conf
ere
nc
e
(
IPE
C
-
Nii
ga
ta
2018
-
ECC
E
As
ia)
,
2018,
pp
.
74
0
-
745,
doi
:
10
.
239
19/IPEC.
2018.
8
507972.
[12]
F.
Guastavi
no,
A.
Darda
no,
“
Li
fe
te
sts
on
twiste
d
pai
rs
in
pre
senc
e
of
par
t
ia
l
d
ischa
rge
s:
infl
u
e
nce
of
the
volt
a
ge
wave
form
,
”
I
EEE
Tr
ansacti
ons
on
Diel
e
ct
ri
cs
and
El
e
ct
rica
l
In
sulati
on
,
vo
l.
19
,
no.
1
,
pp
.
45
-
5
2,
Februa
r
y
201
2
,
doi:
10
.
1109/T
D
EI.
2012.
614850
1.
[13]
P.
W
ang,
C.
Zh
eng,
Y.
L
i,
Y.
Le
i
,
A.
Cavalli
n
i,
“
The
PD
and
Endura
nc
e
Features
of
Ena
m
el
e
d
W
ire
s
at
Shor
t
Repe
titiv
e
Im
pulsive
Volta
g
es,
”
IEEE
Elec
t
rical
Insulati
on
Confe
renc
e
EIC
,
2018,
pp
.
572
-
576,
doi
:
10.
1109/E
IC
.
20
18.
8481043.
[14]
F.
Guastavi
no,
L.
Bria
no
,
F.
Gallesi,
E
.
Torell
o,
“
Corona
resista
nt
insulating
s
y
stems
cha
ra
cteri
z
at
ion
for
lo
w
volt
ag
e
rota
t
ing
m
ac
hine
,
”
IE
EE
Confe
renc
e
on
El
e
ct
rica
l
Insulati
on
and
Dielec
t
ric
Phe
nomen
a
CEIDP
,
2019,
p
p.
694
-
697,
doi
:
10
.
1109/CE
IDP
47102.
2019.
90099
01.
[15]
I.
F.
R
adz
i
,
N
.
H
.
N.
Al
i, N.
R
ameli,
A.
M
.
Ariffi
n,
M.
S.
A.
R
ah
m
an,
M.
R
.
Ahm
ad,
A.
S.
Sa
ll
eh
,
“
Parti
al
Disch
ar
ge
Dete
c
ti
on
Dev
ice
Us
ing
Ultra
so
nic
Sensor
on
Medium
Volta
g
e
XLPE
Cab
le
,
”
Indone
sian
Jou
rnal
of
Elec
trical
Engi
ne
ering
and
Computer
Sc
ie
n
ce
,
vol
.
17
,
no
.
2
,
pp
.
886
-
894
,
2
020
,
doi
:
10
.
115
91/i
jeec
s.v17
.
i2
.
pp886
-
894.
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Mo
deling of
hi
gh freq
ue
ncy
hi
gh
v
olta
ge of
wavef
orms
on l
if
e o
f en
am
el
insu
l
atio
n
(
S
N
ar
asi
m
ha R
ao
)
1339
[16]
A.
Azm
i,
K.
Y.
La
u,
N.
A.
Ahm
ad,
Z.
A.
Mal
ek,
C.
W
.
Ta
n,
“
El
ec
tr
ical
Breakdow
n
and
Chemica
l
Properties
of
Pol
y
prop
y
l
ene
/
Cal
ci
um
Carbon
at
e
Nano
composi
te
s,
”
Indon
esia
n
Journal
of
El
e
ct
rical
Engi
ne
eri
ng
and
Computer
Sci
en
ce
,
vol
.
20
,
no.
2,
pp.
563
-
5
68,
2020
,
doi
:
10
.
11591/i
j
eecs.v2
0.
i2.
pp563
-
568
.
[17]
M.
Florkows
ki
,
B.
Florkows
ka,
P.
Z
y
d
ron,
“
Partial
Discha
rg
e
s
in
Insulat
i
ng
S
y
stems
of
Low
Volta
ge
E
le
c
tric
Motors
Fed
by
Pow
er
El
ec
tron
i
cs
-
Twiste
d
-
Pair
Sam
ple
s
Eva
lua
t
ion,
”
Journal
of
ene
rgies,
vol
.
12,
no.
5,
pp.
1
-
1
9,
2019
,
doi
:
10
.
33
90/e
n12050768.
[18]
A.
Mirza
,
W
.
Chen,
H.
Ngu
y
en
,
Y.
Cao,
A.
M.
B
az
z
i,
“
High
-
Volta
ge
High
-
Freque
n
c
y
T
esti
ng
for
Medium
-
Volta
g
e
Motor
Insulat
io
n
Degra
da
ti
on
,
”
IEE
E
Ene
rgy
Conve
rs
ion
Con
gress
and
Ex
positi
on
ECCE
,
20
18,
pp.
2444
-
24
47,
doi:
10
.
1109/E
C
CE.
2018.
855756
8.
[19]
S.
N.
Rao,
K.
El
anse
r
al
a
tha
n
,
“
Influe
nce
of
PW
M
te
chni
que
on
bre
akdown
in
twiste
d
pai
rs,
”
3rd
Inte
rnatio
nal
Confe
renc
e
on
Condit
ion
Ass
essm
ent
Techniques
in
Elec
tri
cal
Syst
ems
CATCO
N
,
2017,
pp.
427
-
431,
doi:
10.
1109/CATC
ON
.
2017.
8280258.
[20]
F.
Guastavi
no
,
A.
Darda
no,
E.
Tore
l
lo,
“
Mea
s
uring
par
t
ia
l
dis
cha
rge
s
und
er
p
ulsed
voltage
c
ondit
ions,
”
IE
E
E
Tr
ansacti
ons
on
Diel
e
ct
rics
an
d
El
e
ct
rica
l
Insulati
on
,
vo
l.
15
,
no.
6,
pp
.
16
40
-
1648,
Dec
e
m
ber
2008,
doi
:
10.
1109/T
DEI
.
2
008.
4712668.
[21]
K.
Zhou,
T.
Ta
o
,
Y.
C.
Hung,
X.
T.
Li
,
W
.
Zha
o
,
“
The
Eff
ec
t
of
Diffe
ren
t
Rise
T
ime
on
Parti
al
Discha
rge
Act
ivit
y
of
Inte
r
-
Turn
In
sulat
ion
Und
er
Square
W
ave
V
olt
ag
e,
”
Int
ernat
ional
Tr
ansacti
o
ns
on
El
e
ct
ri
cal
Ene
rgy
Syste
ms
,
vol.
23
,
pp
.
1083
-
1094,
2013
,
doi
:
10.
1002
/e
t
ep.
1
639.
[22]
F.
Guastavi
no,
A.
Rat
to,
S.
Squarc
i
a,
E
.
Torello, “El
ec
tro
-
th
erma
l
agi
ng
te
sts on
diffe
ren
t
kinds o
f
ena
m
el
l
ed
wire
s,
”
10th
IE
EE Int
er
nati
onal
Con
fe
re
nce
on
Sol
id
Die
le
c
tric
s
,
2010,
p
p.
1
-
4
,
doi
:
10
.
1
109/ICSD.2010.5568020.
[23]
F.
Guastavi
no,
D.
Cordano,
E.
Tore
l
lo,
L
.
Della
Giovanna
,
“
PD
evol
uti
on
of
conve
nt
iona
l
an
d
cor
ona
resistant
ena
m
el
s,”
IEEE
Confe
renc
e
on
El
e
ct
rica
l
Insul
ati
on
and
Diele
ct
ric
Phe
nomen
a
CEIDP
,
2016,
pp.
259
-
262,
d
oi:
10.
1109/CE
IDP
.
2016.
7785686.
[24]
G.
C.
Mont
ana
ri
,
P.
Ser
i,
R
.
Heb
ner
,
“
T
y
p
e
Of
Suppl
y
W
ave
for
m
,
Parti
a
l
Disch
arg
e
B
eha
vior
And
Li
f
e
Of
Rot
atin
g
Mac
hine
Insula
tion
Sy
st
e
m
s,”
IE
EE
Inte
rnatio
nal
Powe
r
Modulat
or
and
High
Vol
tage
Confe
ren
ce
IPM
HVC
,
2018,
pp.
176
-
179
,
doi
:
10.
1109
/IPMH
VC.2018.
89367
75.
[25]
T.
Hoang,
Y.
V.
Serd
y
uk
,
S.
M.
Gubanski,
“
Mec
hani
sm
s
of
surfac
e
pote
nt
ial
dec
a
y
on
ena
m
el
wir
e
coa
ti
ngs
,
”
IEEE
Tr
ansacti
ons
on
Diel
e
ct
rics
an
d
El
e
ct
rica
l
Insulati
on
,
vo
l.
22
,
no.
6,
pp
.
34
70
-
3480,
Dec
e
m
ber
2015,
doi
:
10.
1109/T
DEI
.
2
015.
005223.
BIOGR
AP
HI
ES OF
A
UTH
ORS
S
Narasimha
R
ao
was
born
in
Srikalaha
sthi
in
1
977.
He
re
ce
iv
e
d
his
B.
Tech
de
gre
e
from
S.V
Coll
ege
of
Eng
i
nee
ring
,
S.V
Un
ive
rsit
y
in
2004
and
M.E
d
egr
e
e
from
Sath
y
ab
ama
Univer
si
t
y
Chenna
i
,
Indi
a
in
2010
.
He
is
a
rese
arc
h
scho
la
r
in
th
e
Depa
rtment
of
E
EE
,
Puducher
r
y
Te
chno
logi
c
al
U
nive
rsit
y
,
Puduc
her
r
y
,
Ind
ia
,
and
his
ar
ea
s
of
in
t
ere
st
are
insulati
on
in
m
otor
winding
s
,
spac
e cha
rge
m
ea
sure
m
ent
,
m
odeling
on
insulation.
Elans
er
alathan
K
was
b
orn
in
Puducher
r
y
in
1971.
He
recei
v
ed
his
B
.
T
ec
h
degr
ee
from
Maha
tma
Gand
hi
Univer
sit
y
,
Kottay
am
in
19
92
,
M.E
degr
ee
from
the
Indian
I
ns
ti
tut
e
of
S
ci
ence,
Bangal
ore
,
Ind
ia
in
19
99
,
and
Ph.D.
d
egr
ee
from
the
I
ndia
n
I
nst
it
ut
e
o
f
Technol
o
g
y
Kanpur,
Indi
a
in
the
y
e
ar
2009
.
He
is
cur
ren
tly
working
as
a
Pr
ofe
ss
or
in
the
Depa
rtment
of
EE
E
,
Puducher
r
y
T
ec
hno
logi
c
al
Univer
sit
y
,
Pu
duche
rr
y
,
India
,
and
his
ar
ea
s
o
f
int
er
est
ar
e
Parti
al di
sch
arg
e
,
m
odel
ing
of ins
ula
ti
on
,
n
ano
d
ielectri
cs,
and
en
er
g
y
s
tora
ge
.
Evaluation Warning : The document was created with Spire.PDF for Python.