Int
ern
at
i
onal
Journ
al of Ele
ctrical
an
d
Co
mput
er
En
gin
eeri
ng
(IJ
E
C
E)
Vo
l.
8
,
No.
6
,
D
ece
m
ber
201
8,
pp. 50
71
~50
79
IS
S
N: 20
88
-
8708
,
DOI: 10
.11
591/
ijece
.
v8
i
6
.
pp50
71
-
50
79
5071
Journ
al h
om
e
page
:
http:
//
ia
es
core
.c
om/
journa
ls
/i
ndex.
ph
p/IJECE
Hilbert
Based T
esting of
ADC Dif
ferenti
al Non
-
lin
earity
U
sing
Wavelet
Tra
ns
for
m Algo
rithms
Emad
A. Aw
ada
El
e
ct
ri
ca
l
and
C
om
pute
r
Engi
n
e
eri
ng,
Appli
ed
S
ci
en
ce
Private
U
nive
rsit
y
,
Jordan
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Dec
25
, 201
7
Re
vised
Ju
l
2
8
,
201
8
Accepte
d
Aug
11
, 201
8
In
te
sting
Mix
e
d
Signal
Devi
ces
such
as
Analog
to
Digit
a
l
and
Digit
al
to
Analog
Converters,
som
e
d
y
na
m
ic
par
amet
ers,
such
as
Diff
er
ent
i
al
Non
-
Li
ne
ari
t
y
and
In
te
gra
l
Non
-
li
n
earit
y
,
ar
e
ver
y
cr
it
ical
to
evalua
t
i
ng
dev
ises
per
form
anc
e
.
How
eve
r,
such
an
aly
s
is
has
bee
n
notori
ous
for
co
m
ple
xity
and
m
assive
compili
ng
proc
ess.
Therefore,
thi
s
rese
arc
h
will
fo
cus
on
te
stin
g
d
y
nami
c
par
am
et
ers
such
as
Diffe
ren
t
ia
l
Non
-
Li
ne
ari
t
y
b
y
sim
ula
ti
n
g
num
ero
us
num
b
ers
of
bi
ts
Anal
og
to
Dig
it
a
l
Co
nver
te
rs
and
te
st
the
outpu
t
signal
s ba
se
on
n
ew
te
sting
a
lgorithm
s of
W
ave
le
t
tra
n
sform
base
d
on
Hilbe
r
t
proc
ess.
Such
a
new
te
st
ing
a
lgo
rit
hm
should
en
hanc
e
the
t
esti
n
g
proc
ess
b
y
using
le
ss
compili
ng
d
at
a
sam
p
l
es
and
prom
pt
t
esti
ng
resul
ts.
I
n
addi
t
ion,
new
te
sting
result
s
will
be
compare
d
with
the
co
nvent
ion
al
te
stin
g
proc
ess
of
Histogram a
lgorithm
s for
accurac
y
and enactment
.
Ke
yw
or
d:
Diff
e
re
ntial
N
on
-
Li
near
it
y
Digital
-
to
-
A
na
log
C
onve
rters
Discrete
Wave
le
t Tran
s
form
Hilbert T
ra
ns
f
or
m
Copyright
©
201
8
Instit
ut
e
o
f Ad
vanc
ed
Engi
n
ee
r
ing
and
S
cienc
e
.
Al
l
rights re
serv
ed
.
Corres
pond
in
g
Aut
h
or
:
Em
ad
A.
Aw
a
da,
Dep
a
rtm
ent o
f El
ect
rical
an
d
Com
pu
te
r
E
ng
i
neer
i
ng,
Applie
d Sci
enc
e
Pr
i
vate
U
nive
rsity
,
Amm
an
, J
or
dan
.
Em
a
il
: e_awada@asu
.du.j
o
1.
INTROD
U
CTION
The
no
ti
on
of
conve
rting
a
na
log
to
dig
it
al
wav
e
f
or
m
(d
igit
iz
ing
)
is
ve
ry
crit
ic
al
ta
sk
for
Di
gital
Sign
al
Process
ing
(
DS
P
)
a
pp
l
ic
at
ion
s
to
ena
ble
gat
her
i
ng,
analy
zi
ng
,
a
nd
inter
po
la
ti
ng
analo
g
data
in
dig
it
al
do
m
ai
n.
H
owe
ver,
this
c
onve
rsion
of
a
nal
og
to
dig
it
al
re
presentat
ion
re
quires
m
ixed
sig
nal
A
nalo
g
t
o
Digital
Converte
rs
(
A
DCs)
,
w
hich
c
om
es
in
m
any
ty
pes
and
s
pe
ed
f
or
fast
an
d
insta
nta
neou
s
dig
it
iz
ing
process.
Th
ough,
to
e
nsure
t
he
pro
pe
r
pe
rfor
m
ance
of
D
SP
dev
i
ces,
thor
ough
te
sti
ng
pa
ram
e
te
rs
are
re
qu
i
r
ed
to
el
i
m
inate
ADC
s
false
repre
sentat
ion
ou
t
puts
.
F
or
e
xam
ple,
pa
ram
et
er
su
ch
as
Diff
e
ren
ti
al
Non
-
Linearit
y
(DNL)
m
us
t
be
chec
ked
f
or
any
ab
norm
al
ity
ou
t
pu
t
c
har
a
ct
erist
ic
s
du
e
t
o
er
r
or
a
dded
or
de
viati
on
from
the
or
i
gin
al
a
nalo
g wa
vefor
m
[1
]
,
[
2].
As
a
res
ult,
suc
h
te
sti
ng
will
ensure
prop
e
r
perform
an
ce
of
a
pp
li
cat
io
ns
as
exp
ect
e
d.
Howe
ver,
at
m
anu
fact
ur
e
le
vel,
te
sti
ng
of
s
uch
de
v
ic
es
ca
n
be
ver
y
le
ng
t
hy
and
c
om
plica
te
d
especial
ly
w
it
h
higher
nu
m
ber
of
bits
AD
Cs
[
2].
F
or
insta
nc
e,
in
[1
]
-
[3
]
,
c
onve
ntion
al
te
sti
ng
of
s
uch
pa
ram
et
ers
requ
ires
la
r
ge
nu
m
ber
of
data
sam
ples
wh
ic
h
m
ake
te
sti
ng
a
nd
c
ompil
ing
process
ver
y
te
di
ous
ta
sk
a
nd
le
ng
t
hy
process
.
T
hat
is,
al
l
ou
t
pu
t
wav
e
f
orm
data
cod
es
need
t
o
be
i
nc
lud
e
d
in
c
ompil
ing
process
and
c
om
par
ed
with
ideal
c
ode
s
for
DNL
.
E
ach
c
od
e
c
on
ta
in
s
a
la
rg
e
nu
m
ber
of
sam
ples
that
fo
rm
ing
a
fixed
le
ng
t
h
of
cod
e
s
known
as
Least
Sign
i
ficant Bi
ts
(LSB
) [2
]
,
[
4].
Ther
e
f
or
e,
an
increase
in
ADC
s
num
ber
of
bits
res
ults
in
c
od
e
incr
eases
by
2
n
(n
is
nu
m
ber
of
ADC
bits)
an
d
inc
re
as
e
num
ber
of
data
sam
ples
m
aking
th
e
te
sti
ng
pr
ocess
ve
ry
le
ng
t
hy
.
I
n
add
it
io
n,
c
onve
ntion
al
m
et
ho
ds
of
te
sti
ng
wa
ve
for
m
cha
racteri
sti
cs
su
c
h
as
FF
T
and
Sin
us
oi
dal
Histo
gr
am
hav
e
been
kn
own
to
include
e
rror
in
te
sti
ng
com
pile
pr
ocess
an
d
est
i
m
a
ti
on
res
ults
[2
]
,
[
6
]
-
[
8].
Fo
r
instance
,
wh
il
e
FFT
is
base
d
on
a
ddit
ive
no
ise
m
od
el
an
d
require
a
la
r
ge
nu
m
ber
of
sa
m
ples
bit
[1
]
,
[
2
]
,
[
9],
[
10
]
,
Sinu
s
oi
dal
His
togram
m
ajo
rity
of
sa
m
ples
colle
ct
e
d
are
local
iz
ed
at
bo
th
en
ds
of
the
Histo
gra
m
to
pr
od
uce
d
la
rg
e
r
er
ror
near
the
peaks
[
2].
The
refor
e
,
de
viati
on
pro
du
ce
d
by
fau
lt
ed
bits
or
exter
nal
er
ror
will
be
include
d
in
est
im
a
ti
ng
co
de
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
87
08
In
t J
Elec
&
C
om
p
En
g,
V
ol.
8
, N
o.
6
,
Dece
m
ber
2
01
8: 50
71
-
50
79
5072
le
ng
t
h.
A
s
a
res
ult,
te
sti
ng
p
r
oc
ess
m
us
t
be
enh
a
nce
d
to
s
horten
te
sti
ng
p
e
r
iod
o
f
ti
m
e
wit
h
im
po
rtance of
high
sensiti
vity
o
f
dat
a analy
sis.
L
arg
e
num
ber
of
r
e
searc
h
es
has
in
vestig
at
e
d
cl
assic
al
te
sti
ng
al
gor
it
h
m
s
to
enh
a
nce
te
sti
ng
al
gorithm
s.
Wh
il
e
so
m
e
wo
r
ks
ha
ve
pe
rfo
r
m
ed
in
the
are
a
of
F
ourier
T
ran
s
f
or
m
and
Sinu
s
oi
dal
Histogra
m
[1
1]
-
[
1
4
]
,
oth
e
rs
ha
ve
inv
e
sti
gated
ne
w
te
sti
ng
al
go
rithm
s
su
c
h
as W
avel
et
transfor
m
s
[1
]
,
[
2
]
,
[
9
]
,
[
1
5]
,
[
1
6
].
Wh
il
e,
I
n
pr
e
vi
ou
s
works,
[
2
]
-
[
5],
D
NL
e
rror
s
for
nu
m
erous
A
DC
de
vic
es
hav
e
bee
n
te
ste
d
us
i
ng
W
avelet
trans
form
,
this
wo
r
k
will
fo
c
us
on
the
be
ne
fit
of
com
bin
ing
Hilbe
rt
Tr
ansfo
rm
and
Wav
el
et
trans
f
or
m
to
sh
ort
en
te
sti
ng
tim
e
sign
ific
antly
by
red
uci
ng
the
total
nu
m
ber
of
requir
ed
colle
ct
ed
da
ta
sa
m
ples.
Tha
t
is,
us
in
g
Hilbert
Transf
or
m
as
a
base
sig
nal
m
od
ule
an
d
Wav
el
et
tra
nsf
o
rm
as
a
da
ta
extracti
ng
pr
ocess
t
o
achieve
faster
test
ing
process
and of
higher
te
sti
ng
se
ns
it
ivi
ty
o
f bit
s erro
r a
naly
sis.
2.
THE
ROTI
CAL
BACKG
ROUN
D
Id
eal
ly
,
to
m
e
asur
e
A
DCs
outp
ut
wa
ve
f
orm
in
te
r
m
of
DN
L
,
the
de
vice
unde
r
te
st
(
AD
Cs
)
out
put
vo
lt
age
range
need
to
be
det
erm
ined
fo
r
st
ep
siz
e
est
i
m
ation
(Le
ast
Sig
nificant
Bi
t
(LSB))
(1)
.
Th
at
is,
the
dev
ic
e
outp
ut
fu
ll
vo
lt
age
ra
ng
e
(Full
Scal
e
Ra
ng
(
FSR)
)
is
div
i
de
d
int
o
e
qu
al
segm
ents
of
volt
age.
Eac
h
div
isi
on
will
co
ntain
a s
pecifi
c ra
ng
e
of
vo
lt
age
values
[
2
1
]
[
0
]
1
21
n
n
V
o
lt
a
g
e
V
o
lt
a
g
e
LS
B
(1)
Wh
e
re
n
is
A
D
C n
um
ber
of
bits
Segm
ents
dev
i
at
ion
can
be
m
easur
e
am
on
g
co
ns
ec
utive
cod
e
s
of
LSB.
Ther
e
f
or
e,
D
NL
ca
n
be
determ
inate
b
y (
2)
[
1
]
[
]
1
V
i
V
i
D
NL
LSB
(2)
3.
TE
STING M
ET
HOD
OL
O
GY
In
t
his
sim
ulatio
n
of
te
sti
ng
AD
Cs
D
NL,
t
est
ing
is
ba
sed
on
ge
ner
at
in
g
dig
it
al
wavef
or
m
cod
es
to
m
easur
es
ou
t
put
co
rr
es
pondi
ng
volt
ages.
As
s
how
n
in
[16],
by
ca
pturin
g
outp
ut
volt
age
c
od
e
,
Hilbert
Transf
or
m
and
Wav
el
et
dec
om
po
sit
ion
will
be
ap
plied
re
sp
ect
ively
to
trans
fer
si
gn
al
and
el
im
inate
extra
unre
qu
ire
d dat
a.
In
t
his
pr
opos
e
d
te
sti
ng
al
gori
thm
s,
Hilbert
Transf
or
m
is
use
d
to
de
fine
t
he
real
a
nd
im
agina
ry
pa
rts
of
a
sig
nal
(3)
.
S
uch
a
proce
ss
will
en
ha
nc
e
the
a
bili
ty
to
detect
f
unct
io
n
peak
th
rou
gh
in
te
rpolat
ion
[
16
]
,
[17
]
. As a
r
e
sul
t,
the co
m
pu
ta
ti
on
of
wa
vefo
rm
sa
m
ple
s can be
base
d on
on
e
p
a
rt
of
div
i
de
c
om
plex
sig
nal
.
z(t) = g
(t)
+
j
()
gt
(
3
)
Howe
ver,
i
n
t
hi
s
wor
k
,
th
e
im
aginar
y
par
t
of
te
ste
d
wav
e
form
ˆ
gt
will
be
c
ollec
te
d
(a
sum
m
a
ti
on
of
in
put
cosine
w
a
ve
a
nd
oth
e
r non
-
i
nput c
om
po
ne
nt
s (no
ise
s
n
) (4).
ˆ
[]
g
t
g
n
n
(
4
)
and the m
odul
at
ed
outp
ut
wa
vefor
m
can be
def
i
ne
as i
n (5)
0
ˆ
s
i
n
(
2
)
g
m
A
f
n
m
(5
)
Yet,
m
od
ulate
d
sig
nal
will
un
de
r
go
f
urt
her
deco
m
po
sit
io
n
an
d
filt
erin
g
process
by
im
ple
m
enting
Di
screte Wa
ve
le
t
Tran
s
f
or
m
(DWT)
.
W
it
h
t
wo
s
pecial
pro
per
ti
es o
f
c
onvoluti
on
a
nd d
own
-
sam
pling
pro
ces
s,
Wav
el
et
tran
s
f
or
m
can
pr
od
uc
e a u
nique alg
or
it
hm
in
analyzing
th
e outp
ut
test
ed
wav
ef
or
m
d
at
a.
Assu
m
ing
a
discrete si
gnal
(
s
n
={s
nk
}), Wa
velet
d
ecom
posit
ion
for
a
ppr
ox
im
a
ti
on
an
d detai
li
ng
co
e
ff
i
ci
ent
can b
e
obta
ined
resp
ect
ively
by
(6)
a
nd
(7) res
pecti
vely
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
& C
om
p
Eng
IS
S
N: 20
88
-
8708
Hilbert B
as
e
d Test
ing of A
D
C Dif
fe
rential
Non
-
li
neari
ty
U
sin
g
W
avelet
Transfor
m…
(
Em
ad A. Aw
ad
a
)
5073
1
,
2
n
j
k
j
n
k
k
s
h
s
(6)
1
,
2
n
j
k
j
n
k
k
d
g
s
(7)
This
process
is don
e
b
y c
onvo
luti
on
a
s s
how
n (8) a
nd (9)
()
(
(
)
*
)
n
j
j
k
nk
k
h
s
h
s
(8)
()
(
(
)
*
)
n
j
j
k
n
k
k
g
s
g
s
(9)
Fo
ll
owed
by
dow
n
-
sam
pling
as in
(10
)
a
nd (11)
1
(
2
)
(
(
)
*
)
nn
s
h
s
(10)
1
(
2)
(
(
)
*
)
nn
d
g
s
(11)
As
a
res
ult,
nu
m
ber
of
proce
s
sed
data
sam
ples
will
be
re
duced
by
half
(
ha
lf
the
ba
nd
widt
h
)
sta
rtin
g
from
the
la
rg
est
ener
gy
le
vel
and
dow
n
int
o
a
su
b
-
le
vel
of
sig
nal
energies
deco
m
po
s
it
ion
[1
8]
-
[
2
3
].
Thi
s
par
ti
cula
r
property
of Wav
el
e
t
trans
form
is
possible by
t
he
adv
a
ntage
o
f
gro
up o
f
filt
er banks o
f
low
-
pa
ss
an
d
High
-
pass
filt
ers.
H
ow
e
ve
r,
ba
sed
on
W
a
velet
f
il
te
r
ban
ks
, whic
h
determ
i
ne
translat
io
n
and
scali
ng p
r
operti
e
s
of
the
wa
ve
form
,
there
are
m
any
t
ypes
of
W
a
velet
uniq
uen
e
ss
known
as
Mothe
r
W
a
velet
s
[9
]
,
[
1
8
]
.
Ther
e
f
or
e
,
the
sel
ect
ion
of
s
pe
ci
fic
ty
pe
fa
m
il
y
W
avelet
can
be
ve
ry
serio
us
ba
se
d
on
Wav
el
et
at
tribu
te
su
ch
as w
i
dth
of fre
qu
e
ncy
window,
decayi
ng, s
ymm
et
ry, r
eg
ul
arit
y, or
th
og
onal
, bio
rth
ogon
al
, etc.
In
t
his w
ork
, Daub
ec
hies W
a
ve
le
ts
(dbxx
)
an
d
Haar
W
a
vele
t
will
be
us
ed
s
ince
they
a
re
w
idely
us
e
d
in
en
gin
ee
rin
g
ap
plica
ti
on
s
a
nd
f
or
m
at
ching
the
pro
per
ti
e
s
of
te
ste
d
si
gnal
data
form
[5
]
.
Yet,
keep
i
ng
i
n
m
ind
the
s
hap
e
of
wav
el
et
fa
m
ilies,
or
th
ogonal
it
y
(su
ch
a
s
Daubec
hies),
a
nd
biorth
ogona
li
ty
(su
ch
as
bi
or
xx
)
of
W
a
velet
tra
ns
f
or
m
will
ha
ve
a
serio
us i
m
pact on test
in
g analy
sis an
d
re
su
lt
s [5],
[24]
-
[
27
]
.
4.
COMP
UTAT
ION
TE
CHNI
QUE
Id
eal
ly
,
A
DCs
ou
tp
ut
wa
vefor
m
is
zero
dev
ia
t
ion
f
ro
m
t
he
ori
gin
al
wa
vefor
m
.
That
is,
base
d
on
sign
al
f
ull
scal
e
range
an
d
A
DC
num
ber
of
bits,
the
outp
ut
dig
it
al
waveform
sh
ou
ld
hav
e
t
he
sam
e
LSB
cod
e
s
th
rou
gh
ou
t
t
he
wav
e
f
orm
vo
lt
age
divi
sion
[2
]
,
[
4].
Yet,
pr
act
ic
al
ly
,
A
DCs
dig
it
al
ou
t
pu
t
sig
na
l
()
xt
con
ta
in
s
er
ror
bu
il
d
into
t
he
ou
t
put
wa
ve
form
(1
2)
due
to
noise
,
hea
t,
an
d
ot
her
e
ff
ect
that
m
ay
cause
wav
e
f
or
m
d
ist
or
ti
on.
(
)
(
)
x
t
x
t
e
(12)
Wh
e
re
()
xt
= O
rigi
nal sig
nal
valu
es, a
nd
e
= error
-
val
ues
That
is,
outp
ut
sign
al
is
a
c
ombinati
on
of
b
ot
h
si
gn
al
data
bi
ts
and
er
ror
bi
ts
.
S
om
e
of
the
se
er
rors
a
re
du
e
to
qu
a
ntiza
ti
on
proces
s
wh
ic
h
de
pends
on
A
DCs
nu
m
ber
of
bits
t
hat
determ
ine
qu
a
ntiza
ti
on
st
ep
siz
e
()
.
As
in
s
how
n
Figure
1,
the
a
nalo
g
in
pu
t
w
a
vefor
m
is
dig
it
iz
ed
by
sam
pli
ng
a
nd
quantiz
at
ion
proc
ess
to
pro
vid
e a
n o
utp
ut
dig
it
al
wavef
orm
as in
Fig
ur
e
2.
C
o
d
e
s
1
1
1
1
1
0
1
0
1
1
0
0
0
1
1
0
1
0
0
0
1
0
0
0
V
-
i
n
Figure
1
.
A
nalog In
pu
t
Sig
na
l
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
87
08
In
t J
Elec
&
C
om
p
En
g,
V
ol.
8
, N
o.
6
,
Dece
m
ber
2
01
8: 50
71
-
50
79
5074
0
0
0
0
0
1
0
1
0
0
1
1
1
0
0
1
0
1
x
0
x
1
x
2
x
3
x
4
x
5
x
n
Figure
2
.
Di
gital
o
ut
pu
t c
odes
As
a
res
ult,
A
DCs
num
ber
of
bits
(n)
will
aff
ect
the
si
ze
of
s
uc
h
a
cod
e
patte
r
n
L
SB
and
any
dev
ia
ti
on
will
resu
lt
in
D
NL
error.
I
n
this
w
ork,
a
real
tim
e
(practi
cal
)
di
gital
wav
e
form
will
be
sim
ula
te
d
as
an
out
pu
t
sig
na
l
based
on
ADC
nu
m
ber
of
bi
ts.
The
ge
ner
at
ed
di
gital
wav
e
form
will
be
m
odulate
d
by
Hi
lbert
Transf
or
m
to
i
den
ti
fy
t
he
rea
l
an
d
im
aginar
y
par
t
of
the
s
ign
al
.
By
capt
ur
i
ng
the
im
ag
inary
par
t,
a
f
ur
t
her
process
will
be
cond
ucted
by
D
W
T
t
o
fet
ch
the
te
ste
d
wav
e
f
or
m
data
by
translat
io
n
an
d
dilat
ion
eff
ect
thr
ough
low
-
pa
ss
an
d
hi
gh
-
pa
ss
filt
ersto
pr
oduce
ap
pro
xim
at
ion
coe
ff
ic
i
ents
(
n
s
)
detai
l
c
oeffici
ents
(
n
d
)
a
s
in
(
13)
a
nd
(14
)
re
sp
ect
ively
.
1
0
1
2
1
,
1
,
1
1
,
0
1
0
1
2
,
0
1
,
1
,
1
1
0
1
2
nn
nn
nn
h
h
h
h
ss
s
h
h
h
h
s
ss
h
h
h
h
(13)
1
0
1
2
1
,
1
1
,
1
1
0
1
2
1
,
0
1
0
1
2
1
,
0
1
0
1
2
1
,
1
1
0
1
2
1
0
1
2
1
,
1
n
n
n
n
n
n
h
h
h
h
s
d
g
g
g
g
s
h
h
h
h
d
g
g
g
g
s
h
h
h
h
g
g
g
g
d
,1
,0
,1
n
n
n
s
s
s
(14)
The
n
dec
om
posed
te
ste
d
sig
na
l
data
by
ta
kin
g
t
he
detai
l
coeffic
ie
nts
(
n
d
)
a
nd
dow
n
sam
ple
by
half
(15) an
d (
16)
1
,
1
1
,
0
1
,
1
1
,
2
1
,
3
1
,
4
1
,
5
1
,
6
1
,
7
(
,
,
,
,
,
,
,
,
)
n
n
n
n
n
n
n
n
n
d
d
d
d
d
d
d
d
d
(15)
1
,
1
1
,
1
1
,
3
1
,
5
1
,
7
(
,
,
,
,
,
)
n
n
n
n
n
d
d
d
d
d
(16)
To
c
om
pu
te
instanta
ne
ous
DN
L
m
easur
e
m
ents
(17),
hi
gh
-
pas
s
coe
ffi
ci
ents
of
sec
ond
le
ve
l
deco
m
po
sit
io
n
will
be
ob
ta
in
ed
to
est
i
m
at
e
the
diff
e
re
nce
betwee
n
the
m
agn
it
udes
of
c
on
s
ecuti
ve
real
cod
es
[1
]
, [
2].
1
,
1
,
1
m
a
x
(
)
1
n
j
n
j
ide
al
dd
D
N
L
n
(17)
wh
e
re
id
e
al
is i
deal
LSB [
1]
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
& C
om
p
Eng
IS
S
N: 20
88
-
8708
Hilbert B
as
e
d Test
ing of A
D
C Dif
fe
rential
Non
-
li
neari
ty
U
sin
g
W
avelet
Transfor
m…
(
Em
ad A. Aw
ad
a
)
5075
5.
SIMULATI
O
N AND
MEA
SURME
NTS
A
range
of
A
na
log
to
Di
gital
Con
ve
rts
was
s
i
m
ulate
d
us
i
ng
Ma
tl
ab
si
m
ulati
on
so
ft
war
e
.
The
ra
ng
e
was
base
d
on
AD
Cs
num
ber
of
bits
(2
-
20
B
it
s)
to
be
te
ste
d
f
o
r
w
orst
case
instanta
neous
DN
L
.
T
he
ge
ne
rated
ou
t
pu
t
si
gnal
was
m
od
ulate
d
by
Hilbe
rt
Transf
or
m
to
ob
ta
in
the
c
ha
racteri
sti
c
of
the
sig
nal
i
n
r
e
al
an
d
i
m
aginar
y pa
rt
as sho
wn in F
igure
3.
Figure
3.
Hilbe
rt Tr
a
ns
form
o
f
real a
nd im
aginar
y
par
t
Upo
n
proce
ssing
t
he
sig
nal
by
Hilbert
Tra
nsfo
rm
,
the
i
m
a
gin
a
ry
par
t
of
the
signa
l
(sine
wav
e
)
wa
s
qu
a
ntize
d an
d conve
rted fo
r
s
a
m
pling
In
te
rleave
as
sho
wn in F
i
gure
4.
Figure
4.
Hilbe
rt Tr
a
nsfo
rm
i
m
aginar
y si
gna
l par
t
,
qua
ntize
d
im
aginar
y pa
rt,
a
nd sam
ple
tim
e
Me
anwhil
e,
th
e
i
m
aginar
y
c
onve
rted
pa
rt
of
the
wa
vefor
m
was
chec
ke
d
by
the
F
ourier
Tra
ns
f
orm
al
gorithm
s to
ve
rify the
sig
nal b
y t
im
e an
d
f
r
equ
e
ncy
do
m
ai
n
as
sho
wn in F
igure
5.
Figure
5.
Hilbe
rt con
ver
te
d si
new
a
ve si
gnal
in tim
e d
om
ai
n
an
d powe
r den
sit
y
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
87
08
In
t J
Elec
&
C
om
p
En
g,
V
ol.
8
, N
o.
6
,
Dece
m
ber
2
01
8: 50
71
-
50
79
5076
Ba
sed
on
su
c
h
al
gorithm
s,
the
quantiz
ed
I
nterleave
si
gn
al
data
f
or
m
is
acqu
i
red
f
or
furt
her
proces
s
of
deco
m
po
sit
ion
to
s
uppres
s
the
wav
e
f
orm
data
sa
m
pl
e
on
m
ulti
le
v
el
as
sp
eci
fied
in
F
ig
ur
e
6
and
7
consecuti
vely
f
or f
irst l
e
vel a
nd sec
ond l
evel
D
WT decom
po
sit
ion p
ro
ce
ss
.
Figure
6.
Disc
r
et
e
W
av
el
et
Tra
ns
f
orm
(
first level dec
om
po
s
it
ion
)
In
first
le
vel
of
D
WT
decom
po
sit
ion
,
the
In
te
rleave
si
gn
al
was
dec
om
po
sed
i
nto
detai
l
an
d
appr
ox
im
at
ion
coef
fici
e
nt
by
sp
li
tt
in
g
the
w
aveform
data
i
nto
hal
f.
Me
a
nwhile
,
the
de
ta
il
coef
fici
ent
fr
om
f
irst
de
com
po
s
it
ion
pr
ocess
was
us
e
d
f
or
a
seco
n
d
r
ound
of
dec
om
po
sit
ion
as
in
F
ig
ure
7.
The
detai
l
par
t
of
the sec
ond
c
oe
ff
ic
ie
ntwa
s
us
e
d
to
esti
m
a
te
f
or instanta
ne
ous DNL.
Figure
7.
Disc
r
et
e
W
av
el
et
Tra
ns
f
or
m
(
Seco
nd level
dec
ompo
sit
io
n)
6.
RESU
LT
S
A
ND D
I
SCUS
S
ION
In
insta
ntane
ous
te
sti
ng
of
dig
it
iz
ed
outp
ut
wav
e
form
thr
ough
pro
posed
te
sti
ng
al
gorithm
s
of
Hilberta
nd
Wa
velet
tran
sf
orm
deco
m
po
sit
i
on
strat
e
gy
,
it
was
cl
ear
tha
t
the
new
al
gorithm
has
show
n
a
n
enh
a
ncem
ent
i
n
both
te
sti
ng
r
esults
an
d
te
ch
niques.
Using
Ma
tl
ab
to
si
m
ulate
var
io
us
nu
m
ber
s
of
bits
AD
Cs
,
and se
ver
al
ty
pe
s of
discrete
Wav
el
et
,
r
es
ults ha
ve
s
how
n
c
lose a
pproxim
a
te
to
co
nventi
onal
test
in
g m
eth
od.
In
t
his
w
ork
,
DN
L
was
the
f
ocus
of
te
sti
ng
an
d
ve
rificat
io
n
in
te
rm
of
ac
cur
acy
with
c
onju
nction
of
sh
ort
ti
m
e
te
st
ing
proces
s.
T
hat
is,
as
wa
ve
le
t
disti
nct
pr
op
e
rtie
s
of
dilat
ion
a
nd
tra
nsl
at
ion
,
a
uniq
ue
ty
pe
wav
el
et
m
us
t
be
us
e
d
to
gi
ve
a
cl
os
e
ap
pr
ox
im
at
e
m
at
ch
of
t
est
ed
wa
ve
form
char
act
erist
ic
s.
F
or
ex
a
m
ple,
gen
e
rated
unde
r
te
st
sine
-
w
aveform
was
te
ste
d
with
sever
al
ty
pes
of
wa
velet
s.
W
it
h
Haar
w
avelet
char
act
e
risti
cs,
as
sho
wn
i
n
F
igure
8,
the
sha
rp
struct
ur
es
edg
e
s
an
d
st
raigh
t
li
ne
of
re
-
a
dju
sti
ng
a
nd
s
hiftin
g
window
did n
ot
g
ive cl
os
e m
at
ch
to test
e
d w
aveform
.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
& C
om
p
Eng
IS
S
N: 20
88
-
8708
Hilbert B
as
e
d Test
ing of A
D
C Dif
fe
rential
Non
-
li
neari
ty
U
sin
g
W
avelet
Transfor
m…
(
Em
ad A. Aw
ad
a
)
5077
Figure
8
.
Haa
r Wavelet
This in
ret
urn,
did
no
t s
how
ve
ry accur
at
e re
su
lt
as show
n
in T
ables
1
-
3.
Me
anwhil
e, th
e D
au
bec
hie
s
fam
i
ly
had
out
perform
ed
the
Haar
Wavel
et
du
e
a
pproxi
m
at
e
si
m
il
arity
of
the
or
i
gina
l
te
ste
d
s
inewav
e
i
n
patte
rn as s
ho
wn in F
ig
ur
e
9.
Figure
9
.
Da
ub
echies fam
il
y
Wav
el
et
That
is,
the
sim
il
arity
in
wa
ve
form
patte
rn
a
ll
ow
s
t
he
Daub
echies
fam
i
ly
wav
el
et
t
o
im
itate
the
un
de
r
te
sti
ng
w
a
vefo
rm
an
d
analy
ze
thr
ough adj
us
t
ed
sh
ifte
d
windows w
it
h
higher
de
te
ct
ion
s
ensiti
vity
.
As
a resu
lt
,
instanta
ne
ous
DN
L
m
easur
e
m
ents
was
obt
ai
ned
an
d
c
om
par
e
d
a
s
s
how
n
i
n
T
a
bles
1
-
3
at
se
ver
al
t
est
ing
fr
e
qu
e
ncies.
Table
1
.
D
NL e
stim
ation
(conv
e
ntio
nal m
et
hod FFT
V
s
. Wavele
t
base
d Hil
ber
t T
ra
ns
f
or
m
1
00
kh
z
HDWT/FF
T
2
Bits
4
Bits
6
Bits
8
Bits
1
0
Bits
1
2
Bits
1
4
Bits
1
6
Bits
1
8
Bits
2
0
Bits
HAAR
0
.74
0
.53
0
.47
0
.39
0
.58
0
.56
0
.13
0
.57
0
.24
0
.38
DB2
0
.56
0
.34
0
.39
0
.37
0
.37
0
.33
0
.31
0
.31
0
.35
0
.26
DB4
0
.53
0
.31
0
.33
0
.31
0
.28
0
.34
0
.29
0
.32
0
.37
0
.22
DB1
0
0
.55
0
.36
0
.36
0
.32
0
.36
0
.39
0
.38
0
.35
0
.36
0
.28
Co
n
v
.
0
.49
0
.44
0
.31
0
.31
0
.33
0
.38
0
.39
0
.32
0
.38
0
.28
Table
2
.
D
NL e
stim
ation
(conv
e
ntio
nal m
eth
od
FFT
V
s
. Wavele
t
base
d Hil
ber
t T
ra
ns
f
or
m
1
50
kh
z
HDWT/FF
T
2
Bits
4
Bits
6
Bits
8
Bits
1
0
Bits
1
2
Bits
1
4
Bits
1
6
Bits
1
8
Bits
2
0
Bits
HAAR
0
.77
0
.39
0
.47
0
.55
0
.52
0
.50
0
.15
0
.53
0
.28
0
.35
DB2
0
.56
0
.32
0
.40
0
.36
0
.39
0
.38
0
.35
0
.33
0
.39
0
.27
DB4
0
.50
0
.30
0
.32
0
.23
0
.24
0
.36
0
.27
0
.23
0
.50
0
.36
DB1
0
0
.58
0
.37
0
.33
0
.30
0
.33
0
.41
0
.40
0
.34
0
.36
0
.28
Co
n
v
.
0
.49
0
.43
0
.33
0
.40
0
.23
0
.31
0
.23
0
.44
0
.46
0
.28
Table
3
.
D
NL e
stim
ation
(conv
e
ntio
nal m
eth
od
FFT
V
s
. Wavele
t
b
ase
d Hil
ber
t T
ra
ns
f
or
m
2
00
kh
z
HDWT/FF
T
2
Bits
4
Bits
6
Bits
8
Bits
1
0
Bits
1
2
Bits
1
4
Bits
1
6
Bits
1
8
Bits
2
0
Bits
HAAR
0
.77
0
.23
0
.47
0
.35
0
.52
0
.50
0
.15
0
.53
0
.28
0
.35
DB2
0
.56
0
.32
0
.40
0
.36
0
.39
0
.38
0
.35
0
.33
0
.39
0
.27
DB4
0
.50
0
.30
0
.30
0
.23
0
.24
0
.36
0
.27
0
.23
0
.50
0
.56
DB1
0
0
.58
0
.36
0
.33
0
.30
0
.33
0
.41
0
.40
0
.34
0
.36
0
.28
Co
n
v
.
0
.58
0
.41
0
.53
0
.40
0
.23
0
.31
0
.20
0
.44
0
.46
0
.28
On
t
he other
hand, usi
ng H
il
ber
t T
ransf
or
m
an
d
W
a
velet
dec
om
po
sit
ion
t
o
el
im
inate
u
nr
equ
i
red
data
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
87
08
In
t J
Elec
&
C
om
p
En
g,
V
ol.
8
, N
o.
6
,
Dece
m
ber
2
01
8: 50
71
-
50
79
5078
and
detect
ion
of
f
unct
io
n
pe
ak
s,
helps
re
duci
ng
the
te
sti
ng
durati
on
ti
m
e
sign
ific
antly
.
That
is,
us
ing
th
e
i
m
aginar
y
wa
vefor
m
par
t
of
Hilbe
rt
Transform
with
the
deco
m
po
sit
ion
process
of
W
a
velet
,
pe
r
m
it
the
reducti
on
of
num
ber
of
c
ompil
ed
data
sam
ples
by
half
at
each
dec
om
posit
ion
le
vel
to
end
-
up
with
ne
arly
on
e
four
t
h
c
ollec
te
d
sam
ple b
it
s for com
pili
ng
proces
s.
7.
CONCL
US
I
O
N
In
t
his
new
w
ork
of
te
sti
ng
D
iffer
e
ntial
N
on
-
Linea
rity
perf
or
m
ance
f
or
A
nalo
g
to
Di
gital
Converts
,
Hilbert
T
ra
ns
f
or
m
was
im
pl
e
m
ented
as
a
n
init
ia
l
m
od
ul
at
ion
te
st
ba
se
.
As
a
res
ult,
a
par
ti
c
ular
da
ta
was
extracte
d
from
the
or
i
gin
al
w
aveform
to
cre
at
e
platf
or
m
for
Discr
et
e
W
a
velet
Algorith
m
s
analy
sis.
Su
ch
a
n
al
gorithm
will
help
re
duci
ng
the
nu
m
ber
of
wa
vefor
m
data
colle
ct
ed
by
the
ad
van
ta
ge
of
bo
t
h
Hilbe
rt
and
Wav
el
et
tra
ns
f
or
m
.
That
is,
l
ess
sam
ples
to
colle
ct
and
store
,
sho
rter
c
om
pi
li
ng
proce
s
s,
an
d
acc
u
rate
error
est
i
m
ation
(DNL) nea
r
t
o
c
onve
ntio
nal esti
m
at
ion
of
Histogram
techn
iq
ue.
ACKN
OWLE
DGE
MENTS
The
aut
hors
ar
e
gr
at
ef
ul
to
Ap
plied
Scie
nce
Pr
ivate
Un
i
ve
rsity
,
Amm
an
-
Jo
r
da
n,
f
or
t
he
fina
ncial
su
pp
or
t
grat
ed t
o
co
ve
r
the
pu
blica
ti
on
fee
of this
pa
pe
r rese
arch arti
cl
es.
REFERE
NCE
S
[1]
T.
Yam
aguc
h
i
a
nd
M.
Som
a,
“
D
y
namic
Te
stin
g
of
AD
Cs
U
sing
W
ave
le
t
Trans
form
”
,
IEE
E
Inte
rnational
T
e
st
Confe
renc
e
,
pp.
379
-
388,
Nov.
1
997.
[2]
C.
Akujuobi
,
E
.
Aw
ada
,
M,
Sa
diku,
and
A.
W
arsa
m
e,
“
W
ave
l
et
-
base
d
d
iffe
r
e
nti
al
nonli
n
ea
r
ity
te
sti
ng
of
m
ixe
d
signal
s
y
s
te
m
A
DCs
”
.
IEEE
Sou
the
ast
Con
,
pp
.
7
6
–
81
,
Mar
200
7.
[3]
P.
Ramos
,
J.
da
Silva
,
D.
Fe
rre
i
ra
and
M.
San
to
s
,
"S
ta
ti
sti
ca
l
l
y
enha
nc
ed
an
al
og
ue
and
m
ixe
d
-
si
gnal
design
and
te
st"
,
IEEE
21st
Inte
rnational
Mi
xe
d
-
Signa
l
Testi
ng
Workshop (
IMSTW)
,
2016,
p
p.
1
-
5.
[4]
Mark
Bake
r
,
“
D
em
y
stif
y
ing
Mix
ed
Signa
l
T
est M
et
hods”.
N
ewn
es,
E
lsevie
r S
ci
e
nce
,
pp.
147
-
237
,
2003
.
[5]
E.
Aw
ada
&
C
.
Akuuobi,
“
DW
T
Te
sting
of
DA
C
Eff
ec
ti
ve
Num
ber
of
Bi
ts”
,
Proceedi
ng
s
of
the
IASTED
Inte
rnational
Co
nfe
renc
e
Circ
uit
and
Syste
m
,
pp
.
145
-
150,
2010
.
[6]
P.
Ramos
,
J.
da
Silva
,
D
.
Ferre
ir
a
and
M
.
Santo
s
,
"S
ta
ti
sti
call
y
enha
nc
ed
an
al
og
ue
and
m
ix
ed
-
si
gnal
design
and
te
st"
,
IEEE
21st
In
te
rnational
Mi
xe
d
-
signal
Testi
ng
Workshop (
IMSTW)
,
pp.
1
-
5
,
2016
.
[7]
Á.
Góm
ez
-
Pau
,
L.
Ba
la
do
,
J.
Figuera
s
,
"M
i
xed
-
signal
te
st
band
guar
din
g
using
digi
tal
l
y
code
d
indi
r
ec
t
m
ea
surem
ent
s"
,
Inte
rnat
ional
Confe
renc
e
on
Synt
h
esis,
Mo
del
ing
,
Ana
ly
si
s
and
Simulati
on
Me
thods
a
nd
Appl
ic
a
ti
ons
to Circuit
D
esign
(
SMACD)
,
pp.
1
-
4
,
2015
.
[8]
E.
Aw
ada
,
"Th
e
Applic
a
ti
on
W
ave
le
t
Tr
ansf
orm
Algorit
hm
in
Te
sti
ng
AD
C
Eff
ec
ti
v
e
Num
ber
of
Bit
s
",
Inte
rnational
Jo
urnal
of
Comput
er
Scienc
e
&
Inf
orm
ati
on
Techn
ology
(
IJCSI
T)
,
Vol
5,
No 5, pp.
161
-
169,
2013
.
[9]
E.
Aw
ada
&
M.
Alom
ari
,
“
Appli
ca
t
ion
of
W
avel
et
Tr
ansform
Anal
y
s
is
to
AD
Cs
Harm
onic
s
Distorti
on,
”
Comput
er
and
Information Sci
en
ce,
Vol
.
6
,
118
-
124,
2013
.
[10]
S.L
.
M.
Hass
an,
N.
Sulai
m
an,
S.S.
Shariff
udin,
T
.
N.T.
Yaa
kub
,
"S
igna
l
-
to
-
noise
Rat
io
Stud
y
o
n
Pipel
ine
d
Fas
t
Fourier
Tr
ansfor
m
Proce
ss
or",
B
ull
etin
o
f
E
lectri
cal
Engi
ne
ering and Inf
orm
ati
cs
,
Vol
7
,
2018
pp.
230
-
235
[11]
B.
Vargha
,
J.
Sc
houkens
,
Y.
Rol
ai
n,
“
Us
ing
red
u
ce
d
-
orde
r
m
odels
in
D/A
conve
rt
er
te
st
ing”
,
IE
E
E
Instrum
ent
ati
o
n
and
Me
asur
eme
nt
Techno
logy
Confe
renc
e
,
pp.
7
01
-
706
,
Ma
y
20
02.
[12]
W
.
Chen
,
Y.
Zhu
,
X
.
Li
u
,
X
.
L
i
,
and
D.
Ou
,"
Co
m
bini
ng
the
hist
ogra
m
m
et
hod
a
nd
the
ult
r
afa
st
segm
ent
ed
m
odel
ide
nti
f
icati
on
of
li
nea
r
ity
err
ors
al
gorit
hm
for
AD
C
li
ne
ari
t
y
t
esti
ng
",
IE
EE
E
uropean
Test
Symposium
,
pp.
1
-
2
,
2016
.
[13]
S.
Max
,
R.
Liggiero
,
"
Enh
anced
AD
C
sine
wave
histogra
m
te
st
",
IEEE
Inte
rnational
Inst
rum
ent
ati
on
and
Me
asur
eme
nt
Te
chnol
ogy
Confe
r
enc
e
Proc
ee
ding
s
,
pp.
1652
–
16
57
,
2015
.
[14]
J.
Dom
ingos,
A.
Ev
ans
,
"
Dig
it
al
pseudora
nd
om
uniform
noise
gen
erators
f
or
AD
C
histog
ram
te
st
",
IE
E
E
Confe
renc
e
on
Design
of
C
ircu
i
ts and
Int
egrated Sy
stems
,
pp.
1
-
6
,
2015
.
[15]
Y.
Min
,
H.
Ki
Kim
,
Y.
Ri
Kang
,
G.
Su
Kim
,
J.
Park
,
S.
Kim
,
"
Des
ign
of
W
ave
le
t
-
B
ase
d
ECG
Dete
ct
or
f
or
Im
pla
nta
bl
e
C
ar
dia
c
Pac
emake
rs
",
IE
EE
Tr
ansac
ti
ons on Biomedical
C
ircui
ts and
Syste
ms
,
2013
,
p
p.
426
-
436
.
[16]
E.
Aw
ada
,
M.
Alom
ari
,
E.
Ra
dwan,
"W
ave
le
t
Esti
m
at
ion
of
THD
Based
on
Hilbe
rt
Tra
nsf
orm
Ext
rac
ti
on"
,
Inte
rnational
Jo
urnal
of Appl
i
ed Engi
ne
ering
R
ese
arch
,
pp
.
6
451
-
6456
,
2016
.
[17]
S.
Abe
y
sek
era
,
"A
n
eff
ic
ie
nt
Hilbe
rt
tr
ansfor
m
int
erp
ola
t
ion
al
gorit
hm
for
pea
k
positi
on
e
stim
at
ion"
,
IE
E
E
Stat
isti
cal Si
gnal
Proce
ss
ing, 11
th
Signal
Proc
essing
Workshop
,
pp
.
417
-
420
,
2001
.
[18]
L.
Peng,
D.
Ji
a
n
-
dong
,
“
Charact
er
isti
c
Ana
l
y
s
i
s
and
Sele
ction
of
W
ave
le
ts
Applic
ab
le
for
Ultra
-
High
-
Spee
d
Protec
ti
on
”
,
IE
E
E
Tr
ansm
ission and
Distributi
on
Confe
ren
ce
,
pp
1
-
5,
2005
.
[19]
A.
Miti
che
,
L.
Miti
che
,
and
H.
Naimi
,
"
Thre
e
le
vel
s
discre
t
e
wave
le
t
tra
nsf
orm
el
li
pti
c
est
i
m
at
ion
for
EC
G
denoi
sing
"
.
4
th
I
nte
rnational
Co
nfe
renc
e
on
Con
trol
Engi
n
ee
ring
&
Information
Technol
ogy
(
CEIT)
,
2016.
pp.
1
-
5.
[20]
S.
Mallat
“
A w
ave
let
tour
of
sign
al
pro
ce
ss
ing
”
A
cade
mic
Press
,
PP
.
220
-
314
,
20
01
.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
& C
om
p
Eng
IS
S
N: 20
88
-
8708
Hilbert B
as
e
d Test
ing of A
D
C Dif
fe
rential
Non
-
li
neari
ty
U
sin
g
W
avelet
Transfor
m…
(
Em
ad A. Aw
ad
a
)
5079
[21]
C.
Esm
era
ld
a,
D.
Poppi
Loday
a
,
"
E
EG
Base
d
Emotion
Monitori
ng
Us
ing
W
ave
le
t
and
L
ea
rning
Ve
ct
o
r
Quanti
z
at
ion
"
,
Proce
ed
ing
of
t
he
El
e
ct
rica
l
E
ngine
ering
Com
pute
r
Sci
en
ce
a
nd
Informatic
s
,
Vo
l
4,
pp.
93
-
10
3
,
2017
.
[22]
E.
Aw
ada
,
"A
naly
s
is
of
SFDR
Us
ing
Pow
er
S
pec
trum
Based
on
W
ave
le
t
Extrac
t
ion",
Inte
rna
ti
onal
Journal
o
f
Appl
ie
d
Eng
inee
ring R
ese
arch
,
p
p.
1256
-
1260
,
2
016
.
[23]
Z.
Bing
y
i
,
C.
Haia
n,
"
Applica
ti
on
of
W
ave
le
t
Multi
-
Resolut
i
on
Anal
y
sis
to
Harm
onic
s
De
te
c
ti
on
Based
on
MA
TL
AB i
n
po
wer
s
y
st
em
",
IE
EE
In
te
rnationa
l
Confe
ren
ce on Aut
omation
and
Logis
ti
cs
,
pp
.
13
7
-
143
,
2017
.
[24]
A.
Redd
y
,
C.
Na
vit
ha
,
K.
Sivani
,
"P
erf
orm
anc
e
E
val
ua
ti
on
of
Adapti
ve
Conti
nuou
s
W
ave
le
t
Tra
ns
form
base
d
Rake
Rec
e
ive
r
for
U
W
B
Sy
st
ems
",
I
nte
rnational
Jou
rnal
of El
e
ct
ri
ca
l
and
Comput
er
Engi
ne
ering
(
IJ
ECE
)
,
Vol
8
,
20
18.
[25]
Y.
Qing
-
Yu,
M.
Long
-
Hu,
"
Fault
Li
ne
De
te
c
ti
o
n
of
Non
-
Eff
ec
t
ive
l
y
E
art
hed
N
eut
ra
l
S
y
stem
Based
onModulus
Maximum
Dete
rm
ini
ng
Polarit
y
"
,
IE
EE
Confe
ren
ce
on
Pow
er
a
nd
Ene
rgy
Eng
ine
e
ring
,
pp.
1
-
4
,
20
09
.
[26]
Y.
Haww
ar,
A.
Rez
a
,
"
Filt
er
ing
(Denoi
sing)
in
t
he
W
ave
let
Tra
n
sform
Do
m
ai
n
",
IEE
E
Int
ernati
o
nal
Confe
renc
e
,
pp.
266
–
269
;
20
00
.
[27]
B.
Silve
rm
an
an
d
J.
Vass
il
ic
os,
W
ave
le
t
:
The
K
e
y
to
Inte
rm
it
t
en
t
Inform
at
ion
(New
York,
NY
:
O
xford
Univer
sit
y
Press
,
2000).
Evaluation Warning : The document was created with Spire.PDF for Python.