I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
pu
t
er
E
ng
ineering
(
I
J
E
CE
)
Vo
l.
15
,
No
.
2
,
A
p
r
il
20
25
,
p
p
.
1
6
5
6
~
1
6
6
9
I
SS
N:
2
0
8
8
-
8
7
0
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijece.
v
15
i
2
.
p
p
1
6
5
6
-
1
6
6
9
1656
J
o
ur
na
l
ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
Im
pro
v
ed conv
o
lutiona
l neural n
et
wo
rk
-
ba
sed bea
ri
ng
f
a
ult
dia
g
no
sis
using
multi
-
pha
se
mo
tor
current
sig
na
ls
H
a
i D
a
ng
H
uu
1
,
Ng
o
c
-
M
y
B
ui
2
,
Va
n
-
P
hu
c
H
o
a
ng
3
,
T
ha
n
g
B
ui Q
uy
3
,
Yen
H
o
a
ng
T
hi
4
1
I
n
st
i
t
u
t
e
o
f
E
l
e
c
t
r
o
n
i
c
s,
A
c
a
d
e
m
y
o
f
M
i
l
i
t
a
r
y
S
c
i
e
n
c
e
a
n
d
T
e
c
h
n
o
l
o
g
y
,
H
a
n
o
i
,
V
i
e
t
n
a
m
2
D
e
p
a
r
t
me
n
t
o
f
Tr
a
i
n
i
n
g
,
A
c
a
d
e
my
o
f
M
i
l
i
t
a
r
y
S
c
i
e
n
c
e
a
n
d
Te
c
h
n
o
l
o
g
y
,
H
a
n
o
i
,
V
i
e
t
n
a
m
3
I
n
st
i
t
u
t
e
o
f
S
y
s
t
e
m
I
n
t
e
g
r
a
t
i
o
n
,
L
e
Q
u
y
D
o
n
T
e
c
h
n
i
c
a
l
U
n
i
v
e
r
si
t
y
,
H
a
n
o
i
,
V
i
e
t
n
a
m
4
F
a
c
u
l
t
y
o
f
R
a
d
i
o
-
E
l
e
c
t
r
o
n
i
c
En
g
i
n
e
e
r
i
n
g
,
Le
Q
u
y
D
o
n
Te
c
h
n
i
c
a
l
U
n
i
v
e
r
si
t
y
,
H
a
n
o
i
,
V
i
e
t
n
a
m
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Sep
1
1
,
2
0
2
4
R
ev
is
ed
No
v
1
5
,
2
0
2
4
Acc
ep
ted
No
v
2
0
,
2
0
2
4
Dia
g
n
o
si
n
g
b
e
a
rin
g
fa
u
lt
s
o
f
th
e
in
d
u
c
ti
o
n
m
o
to
r
is
c
ru
c
ial
fo
r
th
e
m
a
in
ten
a
n
c
e
o
f
ro
tati
n
g
e
lec
tri
c
a
l
m
a
c
h
in
e
s.
Nu
m
e
ro
u
s
m
e
th
o
d
s
h
a
v
e
b
e
e
n
d
e
v
e
lo
p
e
d
a
n
d
p
u
b
li
sh
e
d
fo
r
m
o
n
it
o
ri
n
g
a
n
d
c
las
sify
i
n
g
t
h
e
se
fa
u
lt
s
u
sin
g
se
n
so
r
d
a
ta
su
c
h
a
s
v
ib
ra
ti
o
n
,
a
u
d
io
,
a
n
d
c
u
rre
n
t
sig
n
a
ls.
I
d
e
a
ll
y
,
t
h
e
c
u
rre
n
t
p
h
a
se
s
a
re
b
a
lan
c
e
d
;
h
o
we
v
e
r,
f
a
u
lt
s
d
isru
p
t
t
h
is
s
y
m
m
e
try
,
c
a
u
sin
g
e
a
c
h
p
h
a
se
t
o
re
v
e
a
l
u
n
iq
u
e
d
iag
n
o
stic
d
e
tails.
Co
n
se
q
u
e
n
tl
y
,
st
u
d
ies
th
a
t
re
ly
o
n
a
sin
g
le
p
h
a
se
o
f
t
h
e
c
u
rre
n
t
sig
n
a
l
m
a
y
n
o
t
c
a
p
t
u
re
a
ll
fa
u
lt
-
re
la
ted
c
h
a
ra
c
teristics
.
Re
se
a
rc
h
o
n
m
o
to
r
b
e
a
rin
g
fa
u
lt
d
ia
g
n
o
sis
u
sin
g
tw
o
c
u
rre
n
t
p
h
a
se
s
ty
p
ica
ll
y
e
x
trac
ts
fe
a
tu
re
s
fro
m
e
a
c
h
p
h
a
se
se
p
a
ra
tely
,
a
p
p
ly
i
n
g
m
a
c
h
in
e
lea
rn
in
g
t
o
c
las
sify
t
h
e
fa
u
lt
s.
Cu
rre
n
tl
y
,
n
o
a
p
p
r
o
a
c
h
h
a
s
b
e
e
n
p
ro
p
o
se
d
to
e
x
trac
t
fe
a
tu
re
s
fr
o
m
b
o
th
p
h
a
se
s
sim
u
lt
a
n
e
o
u
sly
.
F
u
r
th
e
rm
o
re
,
th
e
p
ro
p
o
se
d
s
o
lu
t
io
n
s
h
a
v
e
o
n
l
y
b
e
e
n
p
u
b
l
ish
e
d
wit
h
n
o
ise
-
fre
e
d
a
ta.
To
a
d
d
re
ss
th
e
se
c
h
a
ll
e
n
g
e
s,
th
is
p
a
p
e
r
in
tro
d
u
c
e
s
a
n
e
n
h
a
n
c
e
d
s
o
l
u
ti
o
n
th
a
t
imp
ro
v
e
s
th
e
a
c
c
u
ra
c
y
o
f
m
o
t
o
r
b
e
a
rin
g
fa
u
lt
c
las
sifica
ti
o
n
b
a
se
d
o
n
a
n
imp
ro
v
e
d
c
o
n
v
o
l
u
ti
o
n
a
l
n
e
u
ra
l
n
e
two
rk
th
a
t
p
ro
c
e
ss
e
s
c
u
rre
n
t
sig
n
a
ls
fro
m
two
p
h
a
se
s
sim
u
lt
a
n
e
o
u
sly
.
E
x
p
e
rime
n
tal
re
su
lt
s
d
e
m
o
n
stra
te
th
a
t
th
e
p
ro
p
o
se
d
m
e
th
o
d
sig
n
ifi
c
a
n
tl
y
o
u
tp
e
rfo
rm
s
trad
it
io
n
a
l
a
p
p
r
o
a
c
h
e
s,
p
a
rti
c
u
larly
in
sc
e
n
a
rio
s
wh
e
re
t
h
e
sa
m
p
le
sig
n
a
ls
a
re
n
o
ise
-
a
d
d
i
n
g
si
g
n
a
ls
.
F
a
u
lt
c
las
sifica
ti
o
n
a
c
c
u
ra
c
y
o
f
t
h
e
p
ro
p
o
se
d
im
p
ro
v
e
d
c
o
n
v
o
lu
t
io
n
a
l
n
e
u
ra
l
n
e
two
rk
(M
I
-
CNN
)
a
b
o
u
t
9
5
.
1
2
%
wit
h
n
o
ise
-
a
d
d
in
g
sig
n
a
ls
a
t
t
h
e
sig
n
a
l
-
to
-
n
o
ise
ra
ti
o
o
f
2
0
d
B.
K
ey
w
o
r
d
s
:
B
ea
r
in
g
f
au
lt d
iag
n
o
s
is
C
las
s
if
icatio
n
ac
cu
r
ac
y
I
m
p
r
o
v
ed
c
o
n
v
o
lu
tio
n
al
n
eu
r
al
n
etwo
r
k
I
n
d
u
ctio
n
m
o
to
r
Mo
to
r
cu
r
r
en
t sig
n
al
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Van
-
Ph
u
c
Ho
an
g
I
n
s
titu
te
o
f
Sy
s
tem
I
n
teg
r
atio
n
,
L
e
Qu
y
Do
n
T
ec
h
n
ical
Un
iv
er
s
ity
2
3
6
Ho
a
n
g
Qu
o
c
Viet
R
o
ad
,
B
ac
T
u
L
iem
Dis
tr
ict,
Han
o
i,
Vietn
am
E
m
ail:
p
h
u
ch
v
@
lq
d
tu
.
ed
u
.
v
n
1.
I
NT
RO
D
UCT
I
O
N
E
lectr
ical
m
ac
h
in
es
in
g
en
e
r
a
l
an
d
i
n
d
u
ctio
n
m
o
to
r
s
in
p
ar
ticu
lar
ar
e
c
r
u
cial
in
v
ar
i
o
u
s
i
n
d
u
s
tr
ies.
Du
e
to
co
n
tin
u
o
u
s
o
p
er
atio
n
,
th
ese
m
o
to
r
s
ex
p
er
ien
ce
s
ig
n
i
f
ican
t
elec
tr
ical
an
d
m
ec
h
a
n
ical
p
r
ess
u
r
e,
m
ak
in
g
th
em
s
u
s
ce
p
tib
le
to
f
ailu
r
e.
Failu
r
es
ca
n
r
esu
lt
f
r
o
m
th
e
m
o
to
r
its
elf
,
o
p
er
atin
g
c
o
n
d
it
io
n
s
o
r
in
s
tallatio
n
er
r
o
r
s
,
lea
d
in
g
to
r
ed
u
ce
d
m
o
t
o
r
life
s
p
an
an
d
in
c
r
ea
s
ed
p
r
o
d
u
ctio
n
lo
s
s
es
if
n
o
t
d
etec
ted
e
ar
ly
.
C
o
n
s
eq
u
en
tly
,
tech
n
iq
u
es
h
av
e
b
ee
n
d
e
v
elo
p
ed
to
id
en
tify
an
d
ass
ess
f
ailu
r
es,
en
h
an
cin
g
th
e
r
eliab
ilit
y
an
d
av
ailab
ilit
y
o
f
elec
tr
ic
m
o
to
r
s
[
1
]
.
Mo
to
r
f
ailu
r
es
co
m
m
o
n
ly
o
cc
u
r
in
co
m
p
o
n
e
n
ts
s
u
ch
as
b
ea
r
in
g
s
,
s
t
ato
r
s
,
r
o
to
r
s
,
am
o
n
g
o
th
er
s
[
2
]
.
Am
o
n
g
th
ese
co
m
p
o
n
en
ts
,
b
ea
r
i
n
g
f
ailu
r
es
ar
e
th
e
m
o
s
t
p
r
ev
alen
t,
ac
co
u
n
tin
g
f
o
r
o
v
er
4
0
%
o
f
m
o
to
r
f
ailu
r
es.
Acc
o
r
d
in
g
to
r
esear
ch
b
y
Sin
g
an
d
Al
Kaz
za
z
[
3
]
,
b
ea
r
in
g
f
ailu
r
e
is
th
e
p
r
im
ar
y
ca
u
s
e
o
f
m
o
to
r
m
alf
u
n
cti
o
n
s
.
B
ea
r
in
g
s
ar
e
ess
en
tial
f
o
r
th
e
d
u
r
a
b
le
an
d
s
tab
le
o
p
er
atio
n
o
f
m
o
to
r
s
,
s
o
wh
en
th
ey
ar
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
I
mp
r
o
ve
d
co
n
vo
lu
tio
n
a
l n
e
u
r
a
l
n
etw
o
r
k
-
b
a
s
ed
b
ea
r
in
g
fa
u
l
t d
ia
g
n
o
s
is
u
s
in
g
…
(
Ha
i D
a
n
g
Hu
u
)
1657
d
am
ag
ed
,
wh
eth
e
r
th
r
o
u
g
h
b
r
o
k
e
n
r
o
ller
s
o
r
ch
ip
p
ed
i
n
n
er
o
r
o
u
ter
r
in
g
s
,
it
d
ir
ec
tly
im
p
ac
ts
m
o
to
r
p
er
f
o
r
m
an
ce
,
ca
u
s
in
g
im
p
r
o
p
er
o
p
e
r
atio
n
an
d
lo
u
d
n
o
is
es.
T
h
e
v
a
r
io
u
s
ty
p
es
o
f
d
am
a
g
e
in
in
d
u
ctio
n
m
o
to
r
s
ar
e
illu
s
tr
ated
in
Fig
u
r
e
1
[
4
]
.
Fig
u
r
e
1
.
C
lass
if
icatio
n
o
f
co
m
m
o
n
f
a
u
lts
in
in
d
u
ctio
n
m
o
t
o
r
s
[
4
]
Fig
u
r
e
2
d
escr
ib
es
r
o
llin
g
-
ele
m
en
t
b
ea
r
in
g
s
tr
u
ctu
r
e
an
d
f
a
u
lts
,
in
wh
ich
Fig
u
r
e
2
(
a)
illu
s
tr
ates
th
e
s
tr
u
ctu
r
e
o
f
a
ty
p
ical
b
ea
r
in
g
an
d
th
e
d
am
a
g
e
th
at
ca
n
o
c
cu
r
to
its
o
u
te
r
a
n
d
i
n
n
er
r
in
g
s
is
illu
s
tr
ated
in
Fig
u
r
e
2
(
b
)
a
n
d
Fig
u
r
e
2
(
c)
r
e
s
p
ec
tiv
ely
[
5
]
.
A
b
ea
r
in
g
co
n
s
is
ts
o
f
two
r
in
g
s
,
ca
lled
th
e
in
n
er
an
d
o
u
ter
r
in
g
s
,
with
a
s
et
o
f
r
o
llin
g
b
alls
p
o
s
itio
n
ed
o
n
r
o
llin
g
s
u
r
f
ac
es
th
a
t
r
o
tate
ar
o
u
n
d
th
ese
r
i
n
g
s
.
C
o
n
tin
u
o
u
s
s
tr
ess
o
n
th
e
b
ea
r
i
n
g
s
ca
n
lead
to
f
atig
u
e,
ty
p
ically
m
an
i
f
esti
n
g
as
d
am
ag
e
t
o
th
e
in
n
e
r
o
r
o
u
ter
r
in
g
.
T
h
is
d
am
ag
e
r
esu
lts
in
s
m
all
p
iece
s
b
r
ea
k
i
n
g
o
f
f
th
e
b
ea
r
i
n
g
,
a
p
h
e
n
o
m
en
o
n
k
n
o
wn
as
f
lak
in
g
o
r
cr
a
ck
in
g
[
6
]
,
wh
ich
in
tu
r
n
ca
u
s
es
u
n
s
tab
le
b
ea
r
i
n
g
o
p
er
atio
n
.
Sev
er
al
f
ac
t
o
r
s
ca
n
c
o
n
tr
ib
u
te
to
b
ea
r
in
g
f
ailu
r
e,
in
clu
d
in
g
t
h
e
q
u
ality
o
f
th
e
b
ea
r
in
g
its
elf
,
o
p
er
atin
g
in
en
v
ir
o
n
m
en
ts
p
r
o
n
e
to
o
x
id
atio
n
o
r
ch
em
ical
co
r
r
o
s
io
n
an
d
in
s
u
f
f
icien
t
p
er
io
d
ic
m
ain
te
n
an
ce
.
Su
ch
co
n
d
itio
n
s
n
o
t
o
n
ly
im
p
air
t
h
e
s
m
o
o
th
o
p
er
atio
n
b
u
t
also
in
cr
ea
s
e
f
r
ictio
n
,
r
ed
u
cin
g
life
tim
e
o
f
th
e
b
ea
r
in
g
.
T
o
p
r
ev
en
t
p
r
o
d
u
ctio
n
in
ter
r
u
p
tio
n
s
ca
u
s
ed
b
y
e
n
g
i
n
e
f
ailu
r
es,
e
x
ten
d
o
p
er
atin
g
tim
e
a
n
d
o
p
tim
ize
in
v
estme
n
t
ef
f
icien
cy
,
f
au
lt
d
etec
tio
n
an
d
c
o
n
d
itio
n
m
o
n
it
o
r
in
g
ar
e
ess
en
tial.
Fau
lt
d
etec
tio
n
h
elp
s
p
r
ev
e
n
t
u
n
ex
p
ec
te
d
in
ter
r
u
p
tio
n
s
an
d
m
itig
ates
th
e
r
is
k
o
f
s
er
io
u
s
d
am
ag
e
to
th
e
en
tire
p
o
wer
tr
ain
,
wh
ile
c
o
n
d
itio
n
m
o
n
ito
r
in
g
r
ed
u
ce
s
m
ain
ten
an
ce
co
s
ts
an
d
en
h
an
ce
s
en
g
in
e
r
eliab
ilit
y
.
(
a)
(
b
)
(
c)
Fig
u
r
e
2
.
R
o
llin
g
-
elem
e
n
t b
ea
r
in
g
s
tr
u
ctu
r
e
an
d
f
au
lts
(
a)
s
tr
u
ctu
r
e
o
f
a
r
o
llin
g
-
elem
en
t
b
e
ar
in
g
,
(
b
)
o
u
ter
r
ac
e
f
au
lt,
an
d
(
c)
i
n
n
er
r
ac
e
f
a
u
lt
[
5
]
Up
to
n
o
w,
v
ar
i
o
u
s
m
eth
o
d
s
h
av
e
b
ee
n
d
ev
elo
p
e
d
an
d
a
p
p
lied
f
o
r
d
etec
tin
g
a
n
d
d
iag
n
o
s
in
g
m
o
to
r
b
ea
r
in
g
f
au
lts
,
in
clu
d
i
n
g
s
o
u
n
d
an
d
v
ib
r
atio
n
an
aly
s
is
[
7
]
,
elec
tr
o
m
ag
n
etic
f
ield
m
o
n
it
o
r
in
g
[
8
]
an
d
m
o
to
r
cu
r
r
en
t
s
ig
n
al
an
aly
s
is
(
MCS
A)
[
9
]
.
R
esear
ch
er
s
h
av
e
also
in
v
esti
g
ated
f
au
lt
d
iag
n
o
s
is
t
ec
h
n
iq
u
es
b
ased
o
n
o
th
er
m
o
t
o
r
p
h
y
s
ical
q
u
a
n
titi
es,
in
clu
d
in
g
r
o
t
o
r
p
o
s
itio
n
,
r
o
to
r
s
p
ee
d
,
to
r
q
u
e,
p
o
wer
ca
p
ac
ity
,
an
d
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
25
:
1
6
5
6
-
1
6
6
9
1658
tem
p
er
atu
r
e,
in
ad
d
itio
n
to
t
h
ese
s
ig
n
al
-
b
ased
ap
p
r
o
ac
h
es
[
1
0
]
.
Am
o
n
g
th
ese,
v
ib
r
atio
n
an
d
s
o
u
n
d
s
ig
n
al
an
aly
s
is
ar
e
wid
ely
u
s
ed
f
o
r
d
etec
tin
g
m
o
to
r
b
ea
r
in
g
f
a
u
lts
.
Ho
wev
er
,
th
ese
m
eth
o
d
s
r
eq
u
ir
e
th
e
u
s
e
o
f
ex
p
en
s
iv
e
s
en
s
o
r
s
an
d
th
e
p
r
o
p
er
p
lace
m
e
n
t
an
d
in
s
tallatio
n
o
f
th
ese
s
en
s
o
r
s
ca
n
b
e
c
h
allen
g
in
g
d
u
e
to
lim
ited
co
n
s
tr
u
ctio
n
s
p
ac
e.
Ad
d
itio
n
ally
,
th
e
p
r
esen
ce
o
f
n
o
is
e
f
r
o
m
s
u
r
r
o
u
n
d
in
g
d
ev
ice
s
ca
n
in
ter
f
er
e
with
th
e
ac
cu
r
ac
y
o
f
s
o
u
n
d
m
ea
s
u
r
em
en
ts
,
lead
in
g
to
p
o
ten
tial
m
is
d
iag
n
o
s
es
wh
en
u
s
in
g
s
o
u
n
d
s
en
s
o
r
s
to
d
etec
t
b
ea
r
in
g
f
au
lts
.
C
o
m
p
ar
e
d
to
v
ib
r
atio
n
a
n
d
s
o
u
n
d
m
o
n
ito
r
in
g
,
MCS
A
m
eth
o
d
h
as
g
ain
ed
s
ig
n
if
ican
t
atten
tio
n
d
u
e
to
s
ev
er
al
k
ey
ad
v
an
tag
es
.
Firstl
y
,
MCS
A
d
o
es
n
o
t
r
eq
u
ir
e
ad
d
itio
n
al
s
en
s
o
r
s
,
as
it
u
tili
ze
s
th
e
ex
is
tin
g
cu
r
r
en
t
s
ig
n
al
f
r
o
m
th
e
m
o
to
r
co
n
tr
o
ller
,
wh
ic
h
r
ed
u
ce
s
b
o
th
co
s
ts
an
d
s
y
s
tem
co
m
p
lex
ity
.
Ad
d
itio
n
ally
,
th
is
m
eth
o
d
allo
ws
f
o
r
th
e
r
em
o
te
m
o
n
ito
r
in
g
o
f
m
u
ltip
le
m
o
to
r
s
f
r
o
m
a
s
in
g
le
lo
ca
tio
n
b
y
an
aly
zin
g
th
e
cu
r
r
e
n
t
s
ig
n
al
s
u
p
p
lied
to
ea
c
h
m
o
t
o
r
[
1
1
]
.
Fu
r
th
er
m
o
r
e,
MCS
A
is
less
af
f
ec
ted
b
y
am
b
ien
t
n
o
is
e
s
in
ce
it
r
elies
o
n
cu
r
r
en
t sig
n
als f
o
r
d
iag
n
o
s
in
g
b
ea
r
in
g
f
au
lts
,
m
ak
in
g
it a
m
o
r
e
r
eliab
le
o
p
tio
n
in
n
o
is
y
en
v
i
r
o
n
m
en
ts
.
T
y
p
ically
,
b
ea
r
in
g
f
a
u
lt
d
iag
n
o
s
is
u
s
in
g
cu
r
r
en
t
d
ata
with
tr
ad
itio
n
al
m
eth
o
d
s
in
v
o
l
v
es
two
m
ain
s
tep
s
:
f
au
lt
f
ea
tu
r
e
ex
tr
ac
tio
n
an
d
f
au
lt
class
if
icatio
n
.
f
ast
Fo
u
r
ier
tr
an
s
f
o
r
m
(
FF
T
)
[
1
2
]
,
d
is
cr
ete
wav
elet
tr
an
s
f
o
r
m
(
DW
T
)
[
1
2
]
,
em
p
ir
ical
m
o
d
e
d
ec
o
m
p
o
s
itio
n
(
E
MD
)
[
1
3
]
,
l
o
ca
l
m
ea
n
d
ec
o
m
p
o
s
itio
n
(
L
MD
)
[
1
4
]
an
d
v
ar
iatio
n
al
m
o
d
e
d
ec
o
m
p
o
s
itio
n
(
VM
D)
[
1
5
]
ar
e
c
o
m
m
o
n
ly
u
s
ed
f
ea
tu
r
e
ex
tr
ac
tio
n
te
ch
n
iq
u
es.
Fo
r
f
a
u
lt
class
if
icatio
n
,
p
o
p
u
lar
m
ac
h
i
n
e
lear
n
in
g
alg
o
r
ith
m
s
in
clu
d
e
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e
(
SVM)
[
1
6
]
,
b
ac
k
-
p
r
o
p
a
g
atio
n
n
e
u
r
al
n
etwo
r
k
[
1
7
]
,
B
ay
esian
class
if
ier
[
1
8
]
,
k
-
n
ea
r
est
n
eig
h
b
o
r
(
k
-
NN)
[
1
9
]
,
r
an
d
o
m
f
o
r
est
(
R
F)
[
2
0
]
an
d
class
if
icatio
n
an
d
r
eg
r
ess
io
n
tr
ee
(
C
AR
T
)
[
2
1
]
.
T
h
e
ac
cu
r
ac
y
o
f
b
ea
r
in
g
f
au
lt
d
iag
n
o
s
is
u
s
in
g
th
e
ab
o
v
e
a
p
p
r
o
ac
h
es
d
ep
e
n
d
s
o
n
th
e
m
a
n
u
al
s
elec
tio
n
o
f
ex
tr
ac
ted
s
ig
n
al
f
ea
tu
r
es
an
d
tr
ain
in
g
o
f
t
h
e
m
ac
h
in
e
lear
n
in
g
class
if
ier
.
T
o
o
v
e
r
co
m
e
th
is
,
d
ee
p
lear
n
i
n
g
-
b
ased
m
o
to
r
b
ea
r
in
g
f
a
u
lt
d
iag
n
o
s
is
m
eth
o
d
s
,
ty
p
ically
co
n
v
o
l
u
tio
n
al
n
eu
r
a
l
n
etwo
r
k
(
C
NN
)
m
o
d
els,
ar
e
wid
ely
ap
p
lied
[
2
2
]
–
[
2
4
]
.
T
h
ese
ar
e
p
u
b
lis
h
ed
wo
r
k
s
o
n
m
o
to
r
b
ea
r
in
g
f
au
lt
d
iag
n
o
s
is
b
ased
o
n
a
s
in
g
le
p
h
ase
o
f
m
o
to
r
cu
r
r
e
n
t sig
n
al.
I
n
th
eo
r
y
,
wh
en
a
two
-
p
h
ase
elec
tr
ic
m
o
to
r
ex
p
er
ien
ce
s
a
f
au
lt,
th
e
cu
r
r
en
t
in
th
e
t
wo
p
h
ases
b
ec
o
m
es
asy
m
m
etr
ical,
m
ea
n
i
n
g
th
ey
ea
ch
ca
r
r
y
d
is
tin
ct
in
f
o
r
m
atio
n
a
b
o
u
t
th
e
s
y
s
tem
f
au
lt.
T
h
er
ef
o
r
e,
th
e
ac
cu
r
ac
y
o
f
th
e
ab
o
v
e
s
o
lu
tio
n
s
is
n
o
t
h
ig
h
d
u
e
to
m
is
s
in
g
s
y
m
p
to
m
s
o
f
b
ea
r
in
g
f
ailu
r
e.
B
o
th
p
h
ases
o
f
th
e
cu
r
r
en
t
s
ig
n
al
m
u
s
t
b
e
u
s
ed
f
o
r
b
ea
r
i
n
g
f
au
lt
d
iag
n
o
s
is
in
o
r
d
er
to
im
p
r
o
v
e
d
iag
n
o
s
tic
ac
c
u
r
ac
y
an
d
d
ec
r
ea
s
e
m
is
s
ed
d
etec
tio
n
s
.
T
o
d
ate,
th
er
e
ar
e
o
n
ly
a
f
ew
s
tu
d
ies
e
x
p
lo
r
in
g
t
h
e
u
s
e
o
f
two
-
p
h
ase
m
o
to
r
c
u
r
r
en
t
s
ig
n
als
f
o
r
d
iag
n
o
s
in
g
b
ea
r
in
g
f
au
lts
.
Pu
b
lis
h
ed
wo
r
k
s
o
n
d
iag
n
o
s
in
g
m
o
to
r
b
ea
r
in
g
f
a
u
lts
b
ased
o
n
two
p
h
ases
o
f
cu
r
r
en
t
s
ig
n
als
o
n
ly
ex
tr
ac
t
f
e
atu
r
es
o
f
ea
ch
p
h
ase
in
d
iv
id
u
ally
[
5
]
,
[
2
5
]
.
T
h
is
is
n
o
t
s
u
itab
le
f
o
r
d
ia
g
n
o
s
in
g
m
o
to
r
b
ea
r
in
g
f
au
lts
u
s
in
g
m
u
ltip
le
p
h
ases
o
f
cu
r
r
en
t sig
n
als
s
im
u
ltan
eo
u
s
ly
.
T
h
u
s
,
th
is
p
ap
er
p
r
o
p
o
s
es a
n
ew
s
o
lu
tio
n
,
ter
m
ed
m
u
lti
-
in
p
u
t
C
NN
(
MI
-
C
NN)
,
to
o
v
er
co
m
e
th
e
cu
r
r
en
t
d
is
ad
v
a
n
tag
es
o
f
th
e
p
r
ev
i
o
u
s
m
eth
o
d
s
f
o
r
d
etec
tin
g
b
ea
r
in
g
f
ailu
r
es
in
a
m
u
lti
-
p
h
ase
m
o
t
o
r
.
I
n
th
is
m
eth
o
d
,
f
ea
tu
r
e
m
a
p
s
f
r
o
m
b
o
t
h
p
h
ases
o
f
th
e
cu
r
r
en
t
s
ig
n
al
ar
e
ex
tr
a
cted
co
n
c
u
r
r
e
n
tly
th
r
o
u
g
h
two
b
r
an
c
h
es
o
f
th
e
p
r
o
p
o
s
ed
MI
-
C
NN
m
o
d
el.
T
h
es
e
ex
tr
ac
ted
f
ea
tu
r
es
ar
e
th
en
i
n
teg
r
ated
d
u
r
in
g
th
e
f
u
s
io
n
s
tag
e
an
d
s
u
b
s
eq
u
e
n
tly
class
if
ied
b
y
a
s
o
f
tm
ax
class
if
ier
.
Simu
latio
n
s
co
n
d
u
cted
in
v
ar
io
u
s
n
o
is
y
en
v
ir
o
n
m
en
ts
d
em
o
n
s
tr
ate
th
at
th
e
p
r
o
p
o
s
ed
m
eth
o
d
ac
h
iev
es
s
u
p
er
io
r
d
iag
n
o
s
tic
ac
cu
r
ac
y
co
m
p
a
r
ed
to
ex
is
tin
g
s
o
lu
tio
n
s
,
in
clu
d
in
g
th
o
s
e
b
ased
o
n
d
ee
p
lear
n
in
g
an
d
m
ac
h
in
e
lear
n
in
g
with
m
u
lti
-
s
en
s
o
r
s
ig
n
als.
T
h
e
s
u
b
s
e
q
u
en
t
s
ec
tio
n
s
o
f
th
is
ar
ticle
p
r
o
v
id
e
a
th
o
r
o
u
g
h
ex
p
lan
atio
n
o
f
th
e
r
ec
o
m
m
e
n
d
ed
b
e
ar
in
g
f
au
lt
d
iag
n
o
s
is
m
eth
o
d
,
th
e
e
x
p
er
im
e
n
tal
d
at
aset,
v
alid
atio
n
,
a
n
d
d
is
cu
s
s
io
n
.
2.
P
RO
P
O
SE
D
M
E
T
H
O
D
2
.
1
.
B
a
s
ic
CNN
m
o
del st
ruct
ure
f
o
r
dia
g
no
s
ing
elec
t
ric
mo
t
o
r
bea
ring
da
m
a
g
e
T
h
e
b
asic
C
NN
m
o
d
el
s
tr
u
ctu
r
e
f
o
r
d
ia
g
n
o
s
in
g
elec
tr
ic
m
o
to
r
b
ea
r
in
g
f
au
lts
is
d
ep
icted
in
Fig
u
r
e
3
.
I
t
co
m
p
r
is
es
s
ev
er
al
k
ey
co
m
p
o
n
en
ts
:
an
i
n
p
u
t
lay
e
r
(
a
g
r
a
y
s
ca
le
im
ag
e
b
lo
ck
o
f
d
im
en
s
io
n
s
L
1
×
L
2
)
,
f
iv
e
co
n
v
o
l
u
tio
n
al
b
lo
ck
s
,
th
r
ee
p
air
s
o
f
n
o
n
lin
ea
r
an
d
f
u
lly
c
o
n
n
ec
ted
lay
er
s
,
a
s
o
f
tm
ax
la
y
er
,
an
d
an
o
u
tp
u
t
lay
er
.
T
h
e
f
ir
s
t
f
o
u
r
co
n
v
o
lu
ti
o
n
al
b
lo
c
k
s
ea
ch
co
n
s
is
t
o
f
f
o
u
r
lay
er
s
:
a
co
n
v
o
lu
tio
n
al
lay
er
,
a
n
o
r
m
aliza
tio
n
lay
er
,
a
n
o
n
lin
ea
r
ac
tiv
atio
n
lay
er
,
an
d
a
p
o
o
lin
g
lay
er
.
T
h
e
f
if
th
co
n
v
o
l
u
tio
n
al
b
lo
ck
in
cl
u
d
es a
co
n
v
o
lu
tio
n
al
lay
er
an
d
a
n
o
r
m
aliza
tio
n
lay
e
r
.
Pad
d
in
g
is
ap
p
lied
at
th
e
co
n
v
o
lu
tio
n
al
lay
er
s
with
a
s
tr
id
e
o
f
1
×
1
af
ter
ea
ch
m
u
ltip
licatio
n
,
en
s
u
r
in
g
th
at
n
o
in
f
o
r
m
atio
n
is
lo
s
t a
n
d
t
h
at
th
e
im
ag
e
d
im
e
n
s
io
n
s
r
em
ain
u
n
ch
an
g
e
d
.
T
h
e
f
ir
s
t
f
u
lly
co
n
n
ec
ted
lay
er
in
th
e
b
asic
C
NN
m
o
d
el
co
n
tain
s
2
0
0
n
eu
r
o
n
s
,
f
o
llo
w
ed
b
y
th
e
s
ec
o
n
d
f
u
lly
c
o
n
n
ec
te
d
lay
er
with
1
0
0
n
e
u
r
o
n
s
an
d
th
e
f
in
al
f
u
lly
co
n
n
ec
ted
lay
e
r
with
3
n
eu
r
o
n
s
.
T
h
ese
th
r
ee
n
eu
r
o
n
s
co
r
r
esp
o
n
d
to
th
e
b
ea
r
in
g
c
o
n
d
itio
n
lab
els:
0
,
1
,
an
d
2
,
r
ep
r
esen
tin
g
a
n
o
-
f
a
u
lt
b
e
ar
in
g
,
a
n
in
n
er
r
ac
e
b
r
ea
k
b
ea
r
in
g
,
a
n
d
an
o
u
ter
r
a
ce
f
ailu
r
e
b
ea
r
in
g
,
r
esp
ec
tiv
el
y
.
T
h
is
b
asic
C
NN
m
o
d
el,
as d
ep
icted
in
Fig
u
r
e
3
,
o
f
f
er
s
ad
v
an
ta
g
es
s
u
ch
as
lo
w
m
o
d
el
co
m
p
lex
ity
,
r
ap
i
d
tr
ain
in
g
tim
e,
an
d
ef
f
icien
t
i
m
ag
e
class
if
icatio
n
.
Ho
wev
er
,
it
is
lim
i
ted
b
y
its
ab
ilit
y
to
d
iag
n
o
s
e
m
o
to
r
b
ea
r
in
g
f
au
lts
u
s
in
g
o
n
ly
o
n
e
p
h
ase
o
f
th
e
cu
r
r
en
t
s
ig
n
al,
lead
in
g
to
lo
wer
ac
cu
r
ac
y
an
d
r
e
q
u
ir
in
g
a
h
u
g
e
a
m
o
u
n
t
o
f
iter
atio
n
s
f
o
r
th
e
m
o
d
el
to
co
n
v
er
g
e,
p
ar
ticu
lar
ly
in
n
o
is
y
en
v
ir
o
n
m
en
ts
.
T
o
d
iag
n
o
s
e
m
o
to
r
b
e
ar
in
g
f
au
lts
u
s
in
g
b
o
th
p
h
ases
o
f
th
e
m
o
to
r
cu
r
r
en
t
s
im
u
ltan
eo
u
s
ly
,
en
h
a
n
ce
m
en
ts
to
th
e
b
asic CNN m
o
d
el
ar
e
n
ec
ess
ar
y
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
I
mp
r
o
ve
d
co
n
vo
lu
tio
n
a
l n
e
u
r
a
l
n
etw
o
r
k
-
b
a
s
ed
b
ea
r
in
g
fa
u
l
t d
ia
g
n
o
s
is
u
s
in
g
…
(
Ha
i D
a
n
g
Hu
u
)
1659
2
.
2
.
Str
uct
ure
o
f
t
he
pro
po
s
ed
M
I
-
CNN
m
o
del f
o
r
elec
t
ric
mo
t
o
r
bea
ring
f
a
ult
dia
g
no
s
is
T
h
e
p
r
o
p
o
s
ed
MI
-
C
NN
m
o
d
e
l,
illu
s
tr
ated
in
Fig
u
r
e
4
,
is
c
o
m
p
o
s
ed
o
f
f
o
u
r
s
tag
es:
d
ata
co
llectio
n
,
f
ea
tu
r
e
ex
tr
ac
tio
n
,
f
ea
tu
r
e
f
u
s
io
n
,
an
d
class
if
icatio
n
.
T
h
e
d
a
ta
co
llectio
n
an
d
f
ea
tu
r
e
ex
tr
a
ctio
n
s
tag
e
co
n
s
is
t
o
f
two
b
r
an
ch
es:
th
e
lef
t
a
n
d
r
ig
h
t
b
r
an
c
h
es,
wh
ich
ar
e
u
s
ed
to
co
llect
an
d
e
x
tr
ac
t
f
ea
tu
r
es
u
s
in
g
s
ig
n
als
f
r
o
m
Ph
ase
1
an
d
Ph
ase
2
o
f
th
e
m
o
to
r
cu
r
r
en
t,
r
esp
ec
tiv
ely
.
E
ac
h
b
r
an
ch
f
o
llo
ws
an
id
en
tical
s
tr
u
ctu
r
e.
Sig
n
als
f
r
o
m
Ph
ase
1
a
n
d
Ph
ase
2
a
r
e
f
ir
s
t
co
n
v
er
te
d
in
to
g
r
ay
s
ca
le
im
ag
es
o
f
d
im
e
n
s
io
n
s
L
1
×
L
2
.
T
h
ese
im
ag
es
s
er
v
e
as
th
e
i
n
p
u
t
to
th
e
C
NN
m
o
d
el,
wh
ich
co
m
p
r
is
es
f
iv
e
co
n
v
o
lu
tio
n
al
n
o
r
m
aliza
tio
n
r
ec
tifie
d
lin
ea
r
u
n
it
m
ax
-
p
o
o
lin
g
(
C
NR
M
)
b
lo
ck
s
an
d
a
f
u
lly
co
n
n
ec
ted
lay
er
(
F
C
1
)
.
T
h
e
f
ir
s
t
f
o
u
r
C
NR
M
b
lo
ck
s
s
h
ar
e
th
e
s
am
e
s
tr
u
ctu
r
e
as
th
e
co
n
v
o
lu
tio
n
al
b
lo
ck
s
o
f
th
e
b
asic
C
N
N
m
o
d
el
d
escr
ib
ed
in
s
ec
tio
n
2
.
1
.
T
h
e
f
if
th
C
NR
M
b
lo
ck
i
n
clu
d
es
a
co
n
v
o
lu
tio
n
a
l
lay
er
,
a
n
o
r
m
aliza
tio
n
lay
er
,
an
d
a
n
o
n
lin
ea
r
ac
tiv
atio
n
lay
er
.
T
h
e
o
u
tp
u
t
f
r
o
m
th
e
f
ir
s
t
f
u
lly
co
n
n
ec
ted
la
y
er
(
FC
1
)
is
th
e
n
f
ed
in
to
th
e
f
ea
tu
r
e
f
u
s
io
n
b
lo
ck
,
wh
er
e
f
ea
t
u
r
es
ex
tr
ac
ted
f
r
o
m
th
e
two
p
h
ases
o
f
th
e
cu
r
r
en
t
s
ig
n
al
ar
e
co
m
b
in
ed
,
allo
win
g
th
e
MI
-
C
NN
m
o
d
el
to
s
im
u
ltan
eo
u
s
ly
ex
tr
ac
t
f
ea
tu
r
es
f
r
o
m
two
p
h
ases
.
T
h
e
n
u
m
b
er
o
f
n
e
u
r
o
n
s
o
f
th
is
lay
er
is
co
n
ca
ten
ated
b
y
n
e
u
r
o
n
s
o
f
th
e
FC
1
lay
e
r
o
f
th
e
two
b
r
an
ch
es.
T
h
e
f
ea
tu
r
e
class
if
icatio
n
s
tag
e
co
n
s
is
t
s
o
f
two
p
air
s
o
f
r
ec
tifie
d
lin
e
ar
u
n
it
(
R
eL
U
)
an
d
f
u
lly
co
n
n
ec
te
d
lay
er
s
,
f
o
llo
w
ed
b
y
a
s
o
f
tm
ax
lay
er
an
d
an
o
u
tp
u
t
lay
er
.
T
h
is
s
tag
e
clas
s
i
f
ies
th
e
in
p
u
t
im
ag
e
in
to
o
n
e
o
f
th
r
ee
ca
teg
o
r
ies,
lab
eled
0
,
1
,
o
r
2
,
co
r
r
esp
o
n
d
i
n
g
to
d
if
f
er
en
t
b
ea
r
in
g
f
au
lt
c
o
n
d
itio
n
s
,
b
ased
o
n
th
e
p
r
o
b
a
b
ilit
ies
ca
lcu
lated
b
y
th
e
s
o
f
tm
ax
lay
er
,
s
im
ilar
to
th
e
b
asic
C
NN
m
o
d
el.
T
h
e
n
ex
t
s
ec
tio
n
o
f
th
e
p
ap
er
will
p
r
esen
t
th
e
d
ataset
f
o
r
e
x
p
er
im
e
n
tal
v
er
if
icatio
n
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
ef
f
ec
ti
v
en
ess
,
th
e
s
ce
n
ar
io
,
an
d
ex
p
er
im
en
tal
m
eth
o
d
.
Fig
u
r
e
3
.
B
asic CNN m
o
d
el
s
tr
u
ctu
r
e
f
o
r
d
iag
n
o
s
in
g
elec
tr
ic
m
o
to
r
b
ea
r
in
g
d
am
a
g
e
Fig
u
r
e
4
.
Diag
r
a
m
o
f
th
e
p
r
o
p
o
s
ed
MI
-
C
NN
m
o
d
el
s
tr
u
ctu
r
e
f
o
r
d
iag
n
o
s
in
g
elec
tr
ic
m
o
to
r
b
ea
r
in
g
f
au
lts
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
25
:
1
6
5
6
-
1
6
6
9
1660
3.
E
XP
E
R
I
M
E
N
T
A
L
DAT
A
A
ND
M
E
T
H
O
D
3
.
1
.
E
x
perim
ent
a
l da
t
a
W
e
u
s
e
th
e
b
ea
r
in
g
d
ataset
p
r
o
v
id
e
d
b
y
Pad
e
r
b
o
r
n
Un
iv
er
s
ity
[
2
6
]
to
s
im
u
late
th
e
p
r
o
p
o
s
ed
MI
-
C
NN
m
o
d
el,
wh
ich
is
a
lab
o
r
ato
r
y
-
m
ea
s
u
r
e
d
d
ataset
an
d
h
as
b
ee
n
u
s
ed
b
y
m
a
n
y
r
ese
ar
ch
g
r
o
u
p
s
to
test
th
e
p
r
o
p
o
s
ed
b
ea
r
in
g
f
a
u
lt
d
iag
n
o
s
is
s
o
lu
tio
n
(
[
5
]
,
[
2
2
]
,
[
2
3
]
,
[
2
5
]
)
.
T
h
e
p
u
b
lis
h
ed
d
ataset
in
clu
d
es
m
ea
s
u
r
em
en
ts
o
f
v
ar
io
u
s
m
o
to
r
p
ar
am
eter
s
co
r
r
esp
o
n
d
i
n
g
to
3
2
b
ea
r
in
g
co
d
es
o
f
ty
p
e
6
2
0
3
,
with
s
p
ec
if
icatio
n
s
as
d
escr
ib
ed
i
n
T
ab
le
1
.
T
h
ese
b
ea
r
in
g
s
ar
e
ca
teg
o
r
ized
in
to
n
atu
r
ally
d
a
m
ag
ed
,
a
r
tific
ially
d
am
ag
ed
,
a
n
d
u
n
d
am
ag
e
d
g
r
o
u
p
s
,
co
n
s
is
tin
g
o
f
6
u
n
d
am
a
g
ed
b
ea
r
in
g
s
(
K0
0
1
-
K0
0
6
)
,
1
2
b
ea
r
in
g
s
with
o
u
ter
r
in
g
f
a
u
lts
(
KA0
1
,
KA0
3
-
K
A0
9
,
KA1
5
,
KA1
6
,
KA2
2
,
KA3
0
)
,
1
1
b
ea
r
in
g
s
with
in
n
er
r
in
g
f
au
lts
(
KI
0
1
,
KI
0
3
-
KI
0
5
,
K
I
0
7
,
KI
0
8
,
KI
1
4
,
KI
1
6
-
KI
1
8
,
KI
2
1
)
,
a
n
d
3
b
ea
r
in
g
s
with
b
o
t
h
in
n
er
an
d
o
u
ter
r
in
g
f
au
lts
(
KB
2
3
,
KB
2
4
,
KB
2
7
)
.
T
ab
le
1
.
B
ea
r
in
g
p
ar
am
eter
s
o
f
ty
p
e
6
2
0
3
P
a
r
a
me
t
e
r
s
V
a
l
u
e
B
e
a
r
i
n
g
t
y
p
e
D
e
e
p
g
r
o
o
v
e
b
a
l
l
B
e
a
r
i
n
g
d
e
s
i
g
n
a
t
i
o
n
6
2
0
3
D
i
a
me
t
e
r
o
f
i
n
n
e
r
r
a
c
e
2
4
.
0
0
m
m
D
i
a
me
t
e
r
o
f
o
u
t
e
r
r
a
c
e
3
3
.
1
0
m
m
P
i
t
c
h
c
i
r
c
l
e
d
i
a
m
e
t
e
r
2
9
.
0
5
m
m
R
o
l
l
i
n
g
e
l
e
me
n
t
d
i
a
m
e
t
e
r
6
.
7
5
mm
N
u
mb
e
r
o
f
r
o
l
l
i
n
g
e
l
e
me
n
t
s
8
N
o
mi
n
a
l
p
r
e
ss
u
r
e
a
n
g
l
e
0
0
Fig
u
r
e
5
d
is
p
lay
s
th
e
s
ch
em
atic
d
iag
r
am
o
f
th
e
test
s
tan
d
th
a
t
m
ea
s
u
r
es
th
e
p
ar
am
eter
s
o
f
th
e
s
ig
n
als
th
at
d
escr
ib
e
th
e
m
o
to
r
'
s
p
h
y
s
ical
q
u
an
titi
es.
I
t
is
m
ad
e
u
p
o
f
th
e
f
o
llo
win
g
p
ar
ts
:
an
elec
tr
ic
m
o
to
r
(
1
)
,
a
to
r
q
u
e
m
ea
s
u
r
in
g
s
h
af
t
(
2
)
,
a
b
ea
r
in
g
test
m
o
d
u
le
(
3
)
,
a
f
l
y
wh
ee
l
(
4
)
,
an
d
a
lo
ad
m
o
t
o
r
(
5
)
.
B
ea
r
in
g
s
with
d
if
f
er
en
t
f
ailu
r
e
m
o
d
es
a
r
e
m
o
u
n
ted
in
th
e
b
ea
r
in
g
test
m
o
d
u
le
to
g
e
n
er
ate
s
im
u
latio
n
d
ata.
T
h
e
m
o
to
r
(
1
)
is
a
4
2
5
W
p
er
m
an
en
t
m
ag
n
et
s
y
n
ch
r
o
n
o
u
s
m
o
to
r
(
PMSM)
with
a
r
ated
to
r
q
u
e
o
f
T
=
1
.
3
5
Nm
,
a
r
ated
r
o
to
r
s
h
af
t
s
p
ee
d
o
f
n
=
3
0
0
0
r
p
m
,
a
r
ate
d
cu
r
r
en
t
o
f
I
=
2
.
3
A
an
d
th
e
n
u
m
b
er
o
f
p
o
le
p
air
s
o
f
p
=
4
.
I
t
is
o
p
er
ated
b
y
a
f
r
eq
u
e
n
cy
c
o
n
v
e
r
ter
(
KE
B
C
o
m
b
iv
er
t
0
7
F5
E
1
D
-
2
B
0
A)
wit
h
a
c
u
t
-
o
f
f
f
r
e
q
u
en
c
y
o
f
1
6
k
Hz.
T
h
is
m
o
to
r
h
as
4
o
p
er
atin
g
s
tates
co
r
r
esp
o
n
d
i
n
g
to
d
if
f
e
r
en
t
v
alu
es
o
f
r
o
to
r
s
h
af
t
r
o
tatio
n
s
p
ee
d
(
S),
lo
ad
to
r
q
u
e
(
M)
,
r
ad
ial
f
o
r
ce
ac
tin
g
o
n
b
ea
r
in
g
s
(
F).
T
ab
le
2
in
d
icate
s
f
o
u
r
wo
r
k
in
g
co
n
d
itio
n
s
o
f
th
e
m
o
to
r
.
T
h
e
lo
ad
m
o
to
r
(
5
)
is
Siem
en
s
-
Mo
to
r
1
FT7
0
6
2
-
1
AF
7
0
-
1
DG1
.
I
n
th
is
p
ap
er
,
we
u
s
e
two
p
h
ases
o
f
th
e
cu
r
r
en
t
s
ig
n
als
o
f
1
5
b
ea
r
in
g
co
d
es
ar
e
d
ep
icted
in
T
a
b
le
3
,
co
r
r
esp
o
n
d
in
g
to
th
e
wo
r
k
in
g
s
tate
B
o
f
th
e
m
o
to
r
in
T
a
b
le
2
to
d
o
e
x
p
er
im
e
n
t.
Giv
en
a
s
am
p
lin
g
f
r
eq
u
en
cy
o
f
6
4
k
Hz
an
d
a
m
ea
s
u
r
e
m
en
t
d
u
r
atio
n
o
f
4
s
ec
o
n
d
s
,
ea
ch
cu
r
r
en
t
s
ig
n
al
m
ea
s
u
r
em
en
t
c
o
m
p
r
is
es
2
5
6
.
0
0
0
d
ata
p
o
in
ts
(
6
4
.
0
0
0
s
am
p
le
s
/s
ec
o
n
d
×4
s
ec
o
n
d
s
)
.
Fo
r
ea
c
h
b
ea
r
i
n
g
co
d
e,
2
0
m
ea
s
u
r
em
en
ts
ar
e
c
o
n
d
u
cted
,
r
esu
ltin
g
in
a
to
tal
o
f
5
.
1
2
0
.
0
0
0
d
ata
p
o
in
ts
p
er
b
ea
r
in
g
c
o
d
e
(
2
5
6
.
0
0
0
×
2
0
)
.
Sin
ce
ea
ch
b
ea
r
in
g
lab
el
in
T
a
b
le
3
co
n
tain
s
5
b
ea
r
in
g
co
d
es
,
th
e
to
tal
n
u
m
b
er
o
f
d
ata
p
o
in
ts
p
er
b
ea
r
in
g
lab
el
am
o
u
n
ts
to
2
5
.
6
0
0
.
0
0
0
(
5
.
1
2
0
.
0
0
0
×5
)
.
Fig
u
r
e
5
.
T
est s
tan
d
f
o
r
m
ea
s
u
r
in
g
en
g
in
e
p
a
r
am
eter
s
[
2
4
]
T
ab
le
2
.
E
n
g
in
e
wo
r
k
in
g
co
n
d
itio
n
s
O
p
e
r
a
t
i
n
g
st
a
t
e
S
(
R
P
M
)
M
(
N
m)
F
(
N
)
A
1
5
0
0
0
.
1
1
0
0
0
B
9
0
0
0
.
7
1
0
0
0
C
1
5
0
0
0
.
7
4
0
0
D
1
5
0
0
0
.
7
1
0
0
0
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
I
mp
r
o
ve
d
co
n
vo
lu
tio
n
a
l n
e
u
r
a
l
n
etw
o
r
k
-
b
a
s
ed
b
ea
r
in
g
fa
u
l
t d
ia
g
n
o
s
is
u
s
in
g
…
(
Ha
i D
a
n
g
Hu
u
)
1661
T
ab
le
3
.
T
h
e
b
ea
r
in
g
co
d
es u
s
ed
f
o
r
ex
p
er
im
e
n
t
B
e
a
r
i
n
g
c
o
n
d
i
t
i
o
n
C
l
a
s
s
La
b
e
l
B
e
a
r
i
n
g
c
o
d
e
N
o
n
-
f
a
u
l
t
y
(
N
)
0
K0
K
0
0
1
,
K
0
0
2
,
K
0
0
3
,
K
0
0
4
,
K
0
0
5
O
u
t
e
r
r
a
c
e
d
a
m
a
g
e
(
O
)
1
KA
K
A
0
4
,
K
A
1
5
,
K
A
1
6
,
K
A
2
2
,
K
A
3
0
I
n
n
e
r
r
a
c
e
d
a
m
a
g
e
(
I
)
2
KI
K
I
0
4
,
K
I
1
4
,
K
I
1
6
,
K
I
1
8
,
K
I
2
1
3
.
2
.
E
x
perim
ent
a
l m
et
ho
d
T
h
e
d
a
t
a
p
o
i
n
t
s
o
f
b
o
t
h
p
h
a
s
e
s
o
f
t
h
e
m
o
t
o
r
c
u
r
r
e
n
t
s
i
g
n
a
l
w
h
e
n
o
p
e
r
a
t
i
n
g
w
i
t
h
l
a
b
e
l
s
K
0
,
KA
,
a
n
d
K
I
i
n
T
a
b
l
e
3
a
r
e
d
i
v
i
d
e
d
i
n
t
o
g
r
a
y
s
c
a
l
e
i
m
a
g
es
,
e
a
c
h
wi
t
h
a
s
iz
e
o
f
L
1
L
2
p
i
x
e
l
s
.
T
h
e
r
e
f
o
r
e
,
t
h
e
n
u
m
b
e
r
o
f
i
m
a
g
e
s
c
o
r
r
e
s
p
o
n
d
i
n
g
t
o
t
h
e
l
a
b
e
l
s
K0
,
K
A
a
n
d
K
I
i
n
T
a
b
l
e
3
i
s
2
5
.
6
0
0
.
0
0
0
/
(
L
1
L
2
)
i
m
a
g
e
s
.
T
h
e
a
l
l
o
c
a
ti
o
n
o
f
i
n
p
u
t
i
m
a
g
e
s
t
o
t
h
e
M
I
-
C
N
N
m
o
d
e
l
i
s
p
e
r
f
o
r
m
e
d
r
a
n
d
o
m
l
y
a
s
f
o
l
lo
w
s
:
8
0
%
o
f
t
h
e
i
m
a
g
es
a
r
e
as
s
i
g
n
e
d
f
o
r
t
r
a
i
n
i
n
g
a
n
d
t
h
e
r
e
m
a
i
n
i
n
g
2
0
%
a
r
e
r
e
s
e
r
v
e
d
f
o
r
t
e
s
t
i
n
g
a
n
d
e
v
a
l
u
a
t
io
n
.
T
h
e
p
a
r
a
m
e
t
e
r
s
o
f
t
h
e
M
I
-
C
N
N
m
o
d
e
l
f
o
r
t
h
e
e
x
p
e
r
i
m
e
n
t
a
r
e
l
is
t
e
d
i
n
T
a
b
l
e
4
.
T
h
e
m
i
n
i
-
b
a
tc
h
s
i
z
e
i
s
c
h
o
s
e
n
t
o
b
e
1
2
8
.
I
f
t
h
e
v
al
u
e
o
f
t
h
e
m
i
n
i
-
b
a
t
c
h
s
i
z
e
is
t
o
o
s
m
a
l
l
,
t
h
e
n
u
m
b
e
r
o
f
i
t
e
r
a
t
i
o
n
s
wi
l
l
i
n
c
r
e
a
s
e
,
l
e
a
d
i
n
g
to
t
h
e
s
t
a
te
o
f
t
h
e
m
o
d
e
l
g
r
a
d
u
a
l
l
y
m
o
v
i
n
g
f
r
o
m
u
n
d
e
r
f
i
t
t
i
n
g
t
o
o
p
t
i
m
al
a
n
d
t
h
en
o
v
e
r
f
i
t
t
i
n
g
.
O
n
t
h
e
c
o
n
t
r
a
r
y
,
if
t
h
e
m
i
n
i
-
b
a
t
c
h
s
i
ze
is
t
o
o
l
a
r
g
e
,
b
u
f
f
e
r
m
e
m
o
r
y
i
s
n
e
e
d
e
d
t
o
s
t
o
r
e
t
h
e
t
r
a
i
n
i
n
g
d
a
t
a
.
F
u
r
t
h
e
r
m
o
r
e
,
t
o
v
e
r
i
f
y
t
h
e
e
f
f
e
c
t
i
v
e
n
e
s
s
a
n
d
r
e
l
i
a
b
il
i
t
y
o
f
t
h
e
r
e
c
o
m
m
e
n
d
e
d
m
o
d
e
l
,
G
a
u
s
s
i
a
n
n
o
is
e
w
as
a
d
d
e
d
t
o
t
h
e
s
i
g
n
a
ls
u
s
e
d
f
o
r
t
e
s
t
i
n
g
.
I
n
t
h
i
s
p
a
p
e
r
,
t
h
e
p
r
o
p
o
s
e
d
m
e
t
h
o
d
i
s
t
es
te
d
b
y
c
h
a
n
g
i
n
g
t
h
e
c
o
e
f
f
i
ci
e
n
t
s
L
1
,
L
2
(
L
1
L
2
=
4
0
4
0
,
6
0
6
0
,
8
0
8
0
,
a
n
d
1
0
0
1
0
0
)
;
K
(
1
,
5
,
1
0
,
2
0
,
a
n
d
3
0
)
,
F
(
3
3
,
5
5
,
7
7
,
9
9
,
a
n
d
1
1
1
1
)
;
a
n
d
v
a
r
y
i
n
g
t
h
e
v
a
l
u
e
o
f
=
0
.
0
0
1
,
0
.
0
1
,
a
n
d
0
.
1
o
n
m
a
n
y
s
i
g
n
a
l
s
w
i
t
h
d
i
f
f
e
r
e
n
t
s
i
g
n
a
l
-
to
-
n
o
is
e
r
at
i
o
s
,
i
n
c
l
u
d
i
n
g
s
i
g
n
a
l
-
to
-
n
o
is
e
r
at
i
o
(
S
NR
)
=
-
1
0
,
-
5
,
0
,
5
,
1
0
,
1
5
,
a
n
d
2
0
d
B
t
o
s
e
le
c
t
t
h
e
o
p
t
i
m
a
l
p
a
r
a
m
et
e
r
s
e
t
o
f
t
h
e
m
o
d
e
l
b
a
s
e
d
o
n
t
h
e
c
r
i
t
e
r
ia
o
f
f
au
l
t
c
l
ass
i
f
i
ca
t
i
o
n
ac
c
u
r
a
c
y
a
n
d
c
l
a
s
s
i
f
ic
a
t
i
o
n
t
i
m
e
o
f
a
n
i
m
a
g
e
u
s
e
d
f
o
r
m
o
d
e
l
t
es
t
i
n
g
.
A
f
t
e
r
s
e
l
e
ct
i
n
g
t
h
e
o
p
tim
a
l
p
a
r
a
m
e
t
e
r
s
et
,
w
e
c
o
m
p
a
r
e
t
h
e
e
le
c
t
r
i
c
m
o
t
o
r
b
e
a
r
i
n
g
f
a
u
l
t
cl
a
s
s
i
f
ic
a
t
i
o
n
a
c
cu
r
a
c
y
o
f
t
h
e
p
r
o
p
o
s
e
d
s
o
l
u
t
i
o
n
u
s
i
n
g
t
h
e
p
r
o
p
o
s
e
d
M
I
-
C
N
N
m
o
d
e
l
w
i
t
h
t
h
e
b
as
i
c
C
N
N
m
o
d
e
l
a
n
d
c
o
m
p
a
r
e
t
h
e
r
e
s
u
l
ts
o
f
t
h
e
p
r
o
p
o
s
e
d
s
o
l
u
ti
o
n
w
i
t
h
o
t
h
e
r
w
o
r
k
s
u
s
i
n
g
2
-
p
h
a
s
e
d
at
a
s
u
c
h
as
[
5
]
a
n
d
[
2
5
]
u
s
i
n
g
t
h
e
s
a
m
e
d
a
t
a
s
e
t
,
e
x
e
c
u
t
e
d
o
n
t
h
e
s
a
m
e
h
a
r
d
w
a
r
e
p
l
a
t
f
o
r
m
(
I
n
t
e
l
(
R
)
C
o
r
e
i
7
2
.
9
G
H
z
C
P
U;
8
G
B
R
A
M
)
u
s
i
n
g
M
A
T
L
AB
s
o
f
t
w
a
r
e
t
o
h
i
g
h
l
i
g
h
t
t
h
e
e
f
f
e
ct
iv
e
n
e
s
s
o
f
t
h
e
p
r
o
p
o
s
e
d
m
et
h
o
d
.
T
ab
le
4
.
Par
am
eter
s
o
f
th
e
p
r
o
p
o
s
ed
MI
-
C
NN
P
a
r
a
me
t
e
r
V
a
l
u
e
I
n
p
u
t
i
m
a
g
e
s
i
z
e
L1
×
L2
M
i
n
i
b
a
t
c
h
S
i
z
e
1
2
8
I
n
i
t
i
a
l
Le
a
r
n
R
a
t
e
Le
a
r
n
R
a
t
e
D
r
o
p
F
a
c
t
o
r
0
.
1
V
a
l
i
d
a
t
i
o
n
P
a
t
i
e
n
c
e
5
L2
R
e
g
u
l
a
r
i
z
a
t
i
o
n
1e
-
10
Ep
o
c
h
10
Th
e
n
u
m
b
e
r
o
f
k
e
r
n
e
l
s/
t
h
e
k
e
r
n
e
l
si
z
e
p
e
r
c
o
n
v
o
l
u
t
i
o
n
a
l
l
a
y
e
r
s
K
/
F
O
p
t
i
mi
z
e
r
A
d
a
m
4.
E
XP
E
R
I
M
E
N
T
A
L
RE
SUL
T
S AN
D
D
I
SC
USS
I
O
N
4
.
1
.
E
v
a
lua
t
io
n o
f
t
he
f
a
ult
identif
ica
t
io
n a
cc
ura
cy
o
f
t
he
pro
po
s
ed
M
I
-
CN
N
m
o
del w
hen c
ha
ng
ing
t
he
ini
t
ia
l t
ra
ini
ng
ra
t
e
I
n
th
is
ex
p
er
im
en
t,
we
s
elec
ted
th
e
p
r
o
p
o
s
ed
MI
-
C
NN
m
o
d
el
with
a
r
an
d
o
m
s
et
o
f
v
alu
es:
L
1
L
2
(
4
0
×
4
0
)
,
K
(
5
)
,
F
(
7
×
7
)
an
d
v
ar
ied
th
e
lear
n
in
g
r
ate
(
0
.
0
0
1
,
0
.
0
1
,
an
d
0
.
1
)
.
T
h
e
in
p
u
t
d
at
a
co
n
s
is
ted
o
f
two
-
p
h
ase
m
o
to
r
cu
r
r
en
t
s
ig
n
als
with
ad
d
ed
Gau
s
s
ian
n
o
is
e
at
d
if
f
er
en
t
SNR
v
alu
es.
Fi
g
u
r
e
6
s
h
o
ws
th
e
class
if
icatio
n
ac
cu
r
ac
y
o
f
m
o
to
r
b
ea
r
in
g
f
a
u
lts
u
s
in
g
th
e
p
r
o
p
o
s
ed
MI
-
C
NN
m
o
d
el
with
two
-
p
h
ase
m
o
to
r
cu
r
r
en
t
s
ig
n
als
at
a
n
SNR
o
f
2
0
d
B
,
test
ed
at
th
r
ee
d
if
f
er
e
n
t
in
itial
lear
n
in
g
r
ates
with
=
0
.
1
i
n
Fig
u
r
e
6
(
a)
,
=
0
.
0
1
in
Fig
u
r
e
6
(
b
)
,
an
d
=
0
.
0
0
1
in
Fig
u
r
e
6
(
c
)
.
T
h
e
ex
p
er
i
m
en
t
was
r
e
p
ea
ted
u
s
in
g
d
if
f
er
en
t
SNR
v
alu
es a
n
d
th
e
r
esu
lts
o
f
th
e
b
ea
r
in
g
b
r
ea
k
d
o
wn
d
iag
n
o
s
is
ac
cu
r
ac
y
an
d
t
h
e
tim
e
to
class
if
y
a
s
in
g
le
im
ag
e
f
o
r
v
ar
y
in
g
in
itial lea
r
n
in
g
r
ates a
r
e
p
r
esen
ted
in
T
ab
le
5
an
d
T
a
b
le
6
,
r
esp
ec
tiv
ely
.
T
ab
le
5
d
em
o
n
s
tr
ates
th
at
with
th
e
p
r
o
p
o
s
ed
MI
-
C
NN,
th
e
m
o
r
e
th
e
tr
ain
in
g
r
ate
is
r
ed
u
ce
d
,
th
e
m
o
r
e
th
e
f
a
u
lt
class
if
icatio
n
ac
cu
r
ac
y
i
n
cr
ea
s
es,
an
d
th
e
h
ig
h
est
ac
cu
r
ac
y
is
ac
h
iev
e
d
a
t
th
e
r
ate
o
f
0
.
0
0
1
am
o
n
g
th
e
th
r
ee
r
ate
v
alu
es
u
s
ed
f
o
r
ex
p
er
im
en
ts
at
all
ex
p
er
im
en
tal
s
ig
n
als
with
d
if
f
er
e
n
t
SNR
v
alu
es.
T
h
is
alig
n
s
with
th
eo
r
etica
l
ex
p
ec
tatio
n
s
,
wh
er
e
a
lear
n
in
g
r
ate
t
h
at
is
to
o
h
ig
h
m
ay
ca
u
s
e
th
e
m
o
d
el
to
co
n
v
er
g
e
to
o
q
u
ick
ly
to
a
s
u
b
o
p
tim
al
s
o
lu
tio
n
,
wh
ile
a
lear
n
in
g
r
ate
th
at
is
to
o
lo
w
ca
n
ca
u
s
e
th
e
tr
ain
in
g
p
r
o
ce
s
s
to
s
tall.
H
o
wev
er
,
ac
co
r
d
in
g
to
T
ab
le
6
,
d
ec
r
ea
s
in
g
th
e
tr
ain
i
n
g
r
ate
will
r
esu
lt
in
th
e
lo
n
g
er
tim
e
n
ee
d
ed
t
o
class
if
y
a
s
in
g
le
in
p
u
t
im
ag
e
d
u
r
in
g
test
in
g
a
n
d
m
o
d
el
e
v
alu
atio
n
.
T
o
b
alan
ce
b
etwe
en
f
au
lt
class
if
icatio
n
ac
cu
r
ac
y
an
d
ex
ec
u
tio
n
tim
e,
we
ch
o
o
s
e
=
0
.
0
0
1
f
o
r
t
h
e
p
r
o
p
o
s
ed
MI
-
C
NN
m
o
d
el.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
25
:
1
6
5
6
-
1
6
6
9
1662
T
ab
le
5
.
Fau
lt c
lass
if
icatio
n
ac
cu
r
ac
y
with
d
i
f
f
er
en
t i
n
itial lea
r
n
in
g
r
ates
Th
e
i
n
i
t
i
a
l
l
e
a
r
n
i
n
g
r
a
t
e
s
Th
e
f
a
u
l
t
c
l
a
ss
i
f
i
c
a
t
i
o
n
a
c
c
u
r
a
c
y
(
%)
w
i
t
h
t
h
e
n
o
i
se
-
a
d
d
i
n
g
si
g
n
a
l
a
t
t
h
e
d
i
f
f
e
r
e
n
t
S
N
R
2
0
d
B
1
5
d
B
1
0
d
B
5
d
B
0
d
B
-
5
d
B
-
1
0
d
B
0
.
1
3
3
.
6
0
3
3
.
5
5
3
3
.
5
0
3
3
.
5
0
3
3
.
5
0
3
3
.
4
5
3
3
.
3
0
0
.
0
1
8
0
.
6
0
7
6
.
8
3
6
7
.
7
3
5
4
.
6
0
4
4
.
6
8
3
8
.
3
0
3
3
.
5
0
0
.
0
0
1
9
1
.
5
0
8
4
.
4
0
7
1
.
3
0
5
9
.
4
4
5
0
.
2
0
4
6
.
8
0
4
6
.
7
0
T
ab
le
6
.
C
lass
if
icatio
n
tim
e
p
er
im
ag
e
with
d
if
f
er
en
t in
itial l
ea
r
n
in
g
r
ates
Th
e
i
n
i
t
i
a
l
l
e
a
r
n
i
n
g
r
a
t
e
s
N
u
mb
e
r
o
f
i
m
a
g
e
s
t
o
b
e
c
l
a
ss
i
f
i
e
d
Th
e
t
o
t
a
l
w
e
i
g
h
t
(
w
i
t
h
L
1
×
L2
=
40
×
4
0
,
K
=
5
,
F
=
7
×
7)
Ti
me
t
o
c
l
a
s
si
f
y
1
i
ma
g
e
(
ms)
0
.
1
9
6
0
0
5
9
2
4
3
3
.
1
1
±
0
.
3
0
.
0
1
9
6
0
0
5
9
2
4
3
3
.
1
5
±
0
.
3
0
.
0
0
1
9
6
0
0
5
9
2
4
3
3
.
1
9
±
0
.
3
(
a)
(
b
)
(
c)
Fig
u
r
e
6
.
T
h
e
f
au
lt c
lass
if
icatio
n
ac
cu
r
ac
y
o
f
th
e
p
r
o
p
o
s
ed
MI
-
C
NN
m
o
d
el
with
d
if
f
er
e
n
t
in
itial tr
ain
in
g
r
ates,
ex
p
er
im
en
t w
ith
th
e
2
0
d
B
SNR
s
ig
n
al
(
a)
=
0
.
1
,
(
b
)
=
0
.
0
1
,
an
d
(
c)
=
0
.
0
0
1
4
.
2
.
E
v
a
lua
t
io
n
o
f
t
he
f
a
ult
ide
ntif
ica
t
io
n
a
cc
ura
cy
o
f
t
he
pro
po
s
ed
M
I
-
CNN
m
o
del
when
v
a
ry
ing
nu
m
bers o
f
k
er
nels
I
n
th
is
ex
p
er
im
en
t,
we
s
elec
ted
th
e
p
r
o
p
o
s
ed
MI
-
C
NN
m
o
d
el
with
a
r
an
d
o
m
s
et
o
f
v
alu
e
s
:
L
1
×
L
2
(
6
0
×
6
0
)
,
F
(
5
×
5
)
,
(
0
.
0
0
1
)
an
d
v
ar
ied
th
e
K
v
al
u
es
o
f
1
,
5
,
1
0
,
2
0
,
an
d
3
0
.
T
h
e
in
p
u
t
d
at
a
co
n
s
is
ted
o
f
two
-
p
h
ase
m
o
to
r
cu
r
r
e
n
t
s
ig
n
als
with
ad
d
ed
Gau
s
s
ian
n
o
is
e
at
d
if
f
er
en
t
SNR
v
alu
es.
Fig
u
r
e
7
illu
s
tr
ates
th
e
class
if
icatio
n
ac
cu
r
ac
y
o
f
m
o
to
r
b
ea
r
in
g
f
a
u
lts
f
o
r
f
iv
e
d
if
f
er
en
t
k
er
n
el
q
u
an
titi
es
p
er
co
n
v
o
lu
ti
o
n
al
lay
er
u
s
in
g
th
e
p
r
o
p
o
s
ed
MI
-
C
NN
m
o
d
el
u
s
in
g
two
-
p
h
ase
m
o
to
r
cu
r
r
e
n
t
s
ig
n
als
at
an
SNR
o
f
1
5
d
B
with
K
=
1
,
5
,
1
0
,
2
0
,
3
0
in
Fig
u
r
es
7
(
a)
to
7
(
e)
r
esp
ec
tiv
ely
.
T
h
e
e
x
p
er
im
en
t
was
r
ep
ea
ted
u
s
in
g
o
th
er
SNR
v
alu
es
an
d
th
e
r
esu
lts
o
f
class
if
icatio
n
ac
cu
r
a
cy
an
d
th
e
tim
e
ta
k
en
t
o
class
if
y
a
s
in
g
le
im
a
g
e
with
v
ar
y
in
g
k
er
n
el
q
u
a
n
titi
es
ar
e
p
r
esen
ted
in
T
ab
les 7
an
d
8
,
r
esp
ec
tiv
ely
.
Acc
o
r
d
in
g
to
T
a
b
le
7
,
th
e
p
r
o
p
o
s
ed
MI
-
C
NN
m
o
d
el
with
1
k
er
n
el
p
er
co
n
v
o
lu
tio
n
al
la
y
er
h
as
th
e
lo
west
clas
s
if
icatio
n
ac
cu
r
ac
y
.
T
h
e
ac
cu
r
ac
y
is
s
ig
n
if
ican
tly
im
p
r
o
v
ed
if
we
u
s
e
lar
g
er
k
er
n
el
n
u
m
b
er
p
er
co
n
v
o
l
u
tio
n
al
la
y
er
s
u
c
h
as
5
,
1
0
,
2
0
an
d
th
e
ac
cu
r
ac
y
is
m
ax
im
ized
with
a
k
er
n
el
n
u
m
b
er
o
f
3
0
in
al
l
ex
p
er
im
en
tal
s
ig
n
als.
T
h
is
alig
n
s
with
th
eo
r
y
,
wh
ich
s
u
g
g
e
s
ts
th
at
in
cr
ea
s
in
g
th
e
n
u
m
b
er
o
f
k
er
n
els
allo
ws
th
e
C
NN
m
o
d
el
t
o
lear
n
d
ee
p
er
s
ig
n
al
f
ea
tu
r
es.
T
h
er
ef
o
r
e
,
th
e
f
au
lt
class
if
icatio
n
ac
cu
r
ac
y
is
s
ig
n
if
ican
tly
in
cr
ea
s
ed
.
Ho
wev
er
,
as
s
h
o
w
n
in
T
a
b
le
8
,
in
cr
ea
s
in
g
t
h
e
n
u
m
b
er
o
f
k
er
n
els
in
ea
ch
c
o
n
v
o
lu
tio
n
al
la
y
er
also
lead
s
to
a
lar
g
er
n
u
m
b
er
o
f
weig
h
ts
(
4
4
6
8
3
f
o
r
t
h
e
m
o
d
el
with
1
k
er
n
el
v
er
s
u
s
5
2
3
2
0
3
f
o
r
th
e
m
o
d
el
with
3
0
k
er
n
els),
r
esu
ltin
g
in
a
lo
n
g
er
class
if
icatio
n
tim
e
p
er
im
a
g
e
(
2
.
5
8
±
0
.
2
m
s
v
er
s
u
s
6
.
7
9
±
0
.
6
m
s
)
.
T
ab
les
7
an
d
8
s
h
o
w
th
at
u
s
in
g
5
k
er
n
els
p
er
lay
er
y
ield
ed
o
n
ly
a
s
lig
h
tly
lo
wer
class
if
icatio
n
ac
c
u
r
ac
y
co
m
p
ar
ed
to
u
s
in
g
3
0
k
er
n
els (
8
7
.
5
% v
e
r
s
u
s
8
9
.
9
% with
an
SNR
o
f
1
5
d
B
)
.
Ho
wev
er
,
th
e
class
if
icatio
n
tim
e
p
er
im
a
g
e
f
o
r
th
e
m
o
d
el
with
5
k
er
n
els
p
er
co
n
v
o
l
u
tio
n
al
lay
er
was
s
ig
n
if
ican
tly
r
ed
u
ce
d
co
m
p
ar
e
d
to
th
e
m
o
d
el
with
3
0
k
er
n
els,
d
u
e
to
th
e
m
u
ch
lo
w
er
to
tal
n
u
m
b
er
o
f
weig
h
ts
(
6
4
2
0
3
v
er
s
u
s
5
2
3
2
0
3
)
.
T
o
b
al
an
ce
b
etwe
en
f
au
lt
class
if
icatio
n
ac
cu
r
ac
y
an
d
ex
ec
u
tio
n
tim
e,
we
s
elec
t
K
=
5
am
o
n
g
th
e
f
iv
e
k
er
n
el
v
alu
es
u
s
ed
f
o
r
ex
p
er
im
en
tatio
n
f
o
r
th
e
p
r
o
p
o
s
ed
MI
-
C
NN
m
o
d
el.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
I
mp
r
o
ve
d
co
n
vo
lu
tio
n
a
l n
e
u
r
a
l
n
etw
o
r
k
-
b
a
s
ed
b
ea
r
in
g
fa
u
l
t d
ia
g
n
o
s
is
u
s
in
g
…
(
Ha
i D
a
n
g
Hu
u
)
1663
(
a)
(
b
)
(
c)
(
d
)
(
e)
Fig
u
r
e
7
.
Fau
lt c
lass
if
icatio
n
a
cc
u
r
ac
y
o
f
th
e
MI
-
C
NN
m
o
d
e
l w
ith
v
ar
y
in
g
n
u
m
b
er
s
o
f
k
er
n
els,
ex
p
er
im
en
t
with
th
e
1
5
d
B
SNR
s
ig
n
al
(
a)
K=
1
,
(
b
)
K=
5
,
(
c
)
K=
1
0
,
(
d
)
K=
2
0
,
an
d
(
e)
K=
3
0
T
ab
le
7
.
Fau
lt c
lass
if
icatio
n
ac
cu
r
ac
y
with
v
a
r
y
in
g
n
u
m
b
er
s
o
f
k
er
n
els
N
u
mb
e
r
o
f
k
e
r
n
e
l
s
F
a
u
l
t
c
l
a
ssi
f
i
c
a
t
i
o
n
a
c
c
u
r
a
c
y
(
%) w
i
t
h
t
h
e
n
o
i
se
-
a
d
d
i
n
g
s
i
g
n
a
l
a
t
t
h
e
d
i
f
f
e
r
e
n
t
S
N
R
2
0
d
B
1
5
d
B
1
0
d
B
5
d
B
0
d
B
-
5
d
B
-
1
0
d
B
1
7
0
.
3
0
6
5
.
2
0
5
9
.
6
9
5
1
.
2
4
4
6
.
6
8
4
4
.
2
7
4
3
.
6
6
5
9
1
.
7
9
8
7
.
5
0
7
5
.
4
1
7
1
.
3
0
5
0
.
9
1
4
8
.
2
4
4
6
.
4
0
10
9
2
.
3
2
8
9
.
5
9
8
1
.
0
0
7
7
.
2
4
5
3
.
3
5
4
8
.
8
8
4
7
.
5
0
20
9
2
.
8
5
8
9
.
8
0
8
3
.
3
5
8
0
.
1
6
5
4
.
6
8
4
9
.
5
2
4
8
.
4
0
30
9
3
.
5
9
8
9
.
9
0
8
5
.
8
6
8
2
.
5
4
5
8
.
6
2
5
0
.
3
3
4
9
.
6
0
T
ab
le
8
.
C
lass
if
icatio
n
tim
e
p
er
im
ag
e
with
v
a
r
y
in
g
n
u
m
b
er
s
o
f
k
er
n
els
N
u
mb
e
r
o
f
k
e
r
n
e
l
s
/
c
o
n
v
o
l
u
t
i
o
n
a
l
l
a
y
e
r
s
N
u
mb
e
r
o
f
i
m
a
g
e
s
t
o
b
e
c
l
a
ss
i
f
i
e
d
To
t
a
l
w
e
i
g
h
t
(
L1
×
L2
=
6
0
×
6
0
,
F
=
5
×
5)
Ti
me
t
o
c
l
a
s
si
f
y
o
n
e
i
ma
g
e
(
ms)
1
4
2
6
7
4
4
6
8
3
2
.
5
8
±
0
.
2
5
4
2
6
7
6
4
2
0
3
3
.
3
5
±
0
.
3
10
4
2
6
7
9
7
6
0
3
3
.
6
8
±
0
.
3
20
4
2
6
7
3
2
2
4
0
3
5
.
8
2
±
0
.
5
30
4
2
6
7
5
2
3
2
0
3
6
.
7
9
±
0
.
6
4
.
3
.
E
v
a
lua
t
io
n o
f
t
he
f
a
ult
identif
ica
t
io
n a
cc
ura
cy
o
f
t
he
M
I
-
CNN
m
o
del w
hen v
a
ry
ing
kernel
s
izes
I
n
th
is
ex
p
e
r
im
en
t,
we
s
elec
te
d
th
e
p
r
o
p
o
s
ed
MI
-
C
NN
m
o
d
el
with
th
e
f
o
llo
win
g
p
a
r
am
et
er
s
:
L
1
×
L
2
(
8
0
×
8
0
)
,
K
(
1
0
)
,
(
0
.
0
0
1
)
an
d
v
ar
ied
th
e
k
er
n
el
s
izes
F
o
f
3
3
,
5
×
5
,
7
×
7
,
9
×
9
,
an
d
1
1
×
1
1
.
T
h
e
in
p
u
t
d
at
a
im
ag
e
co
n
s
is
ted
o
f
two
-
p
h
as
e
m
o
to
r
cu
r
r
en
t
s
ig
n
als
with
ad
d
ed
Gau
s
s
ian
n
o
is
e
at
d
if
f
er
en
t
SNR
v
alu
es.
Fig
u
r
e
8
illu
s
tr
ates
th
e
cla
s
s
if
icatio
n
ac
cu
r
ac
y
o
f
m
o
to
r
b
ea
r
in
g
f
au
lts
u
s
in
g
th
e
en
h
an
ce
d
C
NN
m
o
d
el
with
f
iv
e
d
if
f
er
en
t
k
er
n
el
s
izes
p
er
co
n
v
o
l
u
tio
n
al
lay
e
r
,
u
s
in
g
two
-
p
h
ase
m
o
t
o
r
c
u
r
r
e
n
t
s
ig
n
als
a
t
an
SNR
o
f
1
0
d
B
with
F
=
3
×
3
,
5
×
5
,
7
×
7
,
9
×
9
,
an
d
1
1
×
1
1
in
Fig
u
r
es
8
(
a)
to
8
(
e)
r
es
p
ec
tiv
ely
.
T
h
e
ex
p
e
r
im
en
t
wa
s
r
ep
ea
t
e
d
with
o
th
er
SNR
v
alu
es,
an
d
th
e
r
esu
lts
f
o
r
class
if
icatio
n
ac
c
u
r
ac
y
an
d
th
e
tim
e
to
class
if
y
a
s
in
g
le
im
ag
e
with
d
if
f
er
en
t
k
er
n
el
s
izes a
r
e
p
r
ese
n
ted
in
T
ab
les 9
an
d
1
0
,
r
esp
e
ctiv
ely
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
25
:
1
6
5
6
-
1
6
6
9
1664
T
ab
les
9
an
d
1
0
d
em
o
n
s
tr
ate
th
at,
with
th
e
p
r
o
p
o
s
ed
m
o
d
el,
in
cr
ea
s
in
g
th
e
k
er
n
el
s
ize
n
o
t
o
n
ly
r
ed
u
ce
s
th
e
m
alf
u
n
ctio
n
id
e
n
tific
atio
n
ac
cu
r
ac
y
b
u
t
also
in
cr
ea
s
es
th
e
to
tal
weig
h
t
o
f
th
e
m
o
d
el,
lead
in
g
to
a
n
in
cr
ea
s
e
in
th
e
tim
e
r
eq
u
ir
ed
to
class
if
y
an
im
ag
e
in
all
ex
p
er
im
en
tal
s
ig
n
als.
Acc
o
r
d
in
g
to
T
ab
les
9
an
d
1
0
,
we
ch
o
s
e
t
h
e
k
er
n
el
s
ize
in
e
ac
h
co
n
v
o
l
u
tio
n
al
la
y
er
as
3
×
3
f
o
r
th
e
s
u
g
g
ested
MI
-
C
NN
m
o
d
el
t
o
r
ea
c
h
t
h
e
h
ig
h
est
ac
cu
r
ac
y
an
d
en
s
u
r
e
th
e
f
astes
t
im
ag
e
class
if
icatio
n
tim
e
co
m
p
ar
ed
to
th
e
r
em
a
in
in
g
k
e
r
n
el
s
izes.
T
h
is
alig
n
s
with
th
eo
r
y
,
wh
ich
s
u
g
g
ests
th
at
ch
o
o
s
in
g
a
s
m
all
k
er
n
el
s
ize
will
ex
tr
ac
t
h
ig
h
ly
lo
ca
l
f
ea
tu
r
es,
d
etec
t sm
all
f
ea
tu
r
es,
ex
tr
ac
t d
iv
er
s
e
f
ea
tu
r
es,
b
e
u
s
ef
u
l f
o
r
t
h
e
f
o
llo
win
g
lay
er
s
,
an
d
s
h
ar
e
weig
h
ts
well.
T
ab
le
9
.
Fau
lt c
lass
if
icatio
n
ac
cu
r
ac
y
with
v
a
r
y
in
g
k
er
n
el
s
izes
K
e
r
n
e
l
si
z
e
/
c
o
n
v
o
l
u
t
i
o
n
a
l
l
a
y
e
r
s
F
a
u
l
t
c
l
a
ssi
f
i
c
a
t
i
o
n
a
c
c
u
r
a
c
y
(
%) w
i
t
h
n
o
i
s
e
-
a
d
d
i
n
g
s
i
g
n
a
l
a
t
d
i
f
f
e
r
e
n
t
S
N
R
v
a
l
u
e
s
2
0
d
B
1
5
d
B
1
0
d
B
5
d
B
0
d
B
-
5
d
B
-
1
0
d
B
3
×
3
9
5
.
7
2
9
1
.
1
2
8
7
.
6
0
7
3
.
3
4
6
2
.
3
5
5
7
.
9
2
5
2
.
1
6
5
×
5
9
4
.
7
2
8
8
.
8
9
8
6
.
7
0
7
2
.
2
8
6
0
.
2
3
5
6
.
2
8
5
1
.
2
1
7
×
7
9
3
.
5
1
8
5
.
3
0
8
0
.
8
0
7
0
.
1
0
5
9
.
2
0
5
5
.
7
8
5
0
.
8
7
9
×
9
9
1
.
1
4
8
4
.
4
8
7
9
.
1
0
6
8
.
8
2
5
8
.
9
0
5
4
.
3
2
5
0
.
3
3
11
×
11
8
9
.
3
5
8
2
.
2
3
7
5
.
9
0
6
6
.
4
7
5
7
.
5
7
5
3
.
6
6
4
8
.
9
0
T
ab
le
1
0
.
C
lass
if
icatio
n
tim
e
p
er
im
ag
e
with
v
ar
y
in
g
k
er
n
el
s
izes
K
e
r
n
e
l
si
z
e
/
c
o
n
v
o
l
u
t
i
o
n
a
l
l
a
y
e
r
s
N
u
mb
e
r
o
f
i
m
a
g
e
s
t
o
b
e
c
l
a
ss
i
f
i
e
d
To
t
a
l
w
e
i
g
h
t
(
w
i
t
h
L
1
×
L2
=
80
×
8
0
,
K
=
1
0
)
T
i
me
t
o
c
l
a
s
si
f
y
o
n
e
i
ma
g
e
(
ms)
3
3
2
4
0
0
1
4
8
4
8
3
3
.
6
4
±
0
.
3
5
5
2
4
0
0
1
6
1
6
0
3
3
.
7
5
±
0
.
3
7
7
2
4
0
0
1
8
1
2
8
3
3
.
9
2
±
0
.
3
9
9
2
4
0
0
2
0
7
5
2
3
3
.
9
6
±
0
.
5
11
11
2
4
0
0
2
4
0
3
2
3
3
.
9
9
±
0
.
6
(
a)
(
b
)
(
c)
(d
)
(e
)
Fig
u
r
e
8
.
Fau
lt c
lass
if
icatio
n
a
cc
u
r
ac
y
o
f
th
e
p
r
o
p
o
s
ed
MI
-
C
NN
m
o
d
el
with
v
ar
y
in
g
k
er
n
el
s
izes,
ev
alu
atio
n
with
th
e
1
0
d
B
SNR
s
ig
n
al
(
a)
F=3
×3
,
(
b
)
5
×5
,
(
c)
F=7
×
7
,
(
d
)
9
×9
,
an
d
(
e)
F=
11
×
11
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
I
mp
r
o
ve
d
co
n
vo
lu
tio
n
a
l n
e
u
r
a
l
n
etw
o
r
k
-
b
a
s
ed
b
ea
r
in
g
fa
u
l
t d
ia
g
n
o
s
is
u
s
in
g
…
(
Ha
i D
a
n
g
Hu
u
)
1665
4
.
4
.
E
v
a
l
u
a
t
i
o
n
o
f
t
h
e
f
a
ul
t
i
de
n
t
if
i
c
a
t
i
o
n
a
c
c
ur
a
c
y
o
f
t
h
e
M
I
-
C
N
N
mo
d
e
l
w
h
en
v
a
r
y
in
g
i
np
u
t
d
a
t
a
s
i
z
es
I
n
th
is
ex
p
er
im
e
n
t,
we
s
elec
ted
th
e
C
NN
m
o
d
el
with
F
=
11
×
1
1
,
K
=
1
,
=0
.
0
0
1
an
d
v
ar
y
in
g
th
e
L1
×
L
2
v
alu
es o
f
4
0
×
4
0
,
6
0
×
6
0
,
8
0
×
8
0
,
an
d
1
0
0
×
1
0
0
with
in
p
u
t d
a
ta
co
n
s
is
tin
g
o
f
two
-
p
h
ase
m
o
to
r
cu
r
r
e
n
t
s
ig
n
als
with
ad
d
ed
Gau
s
s
ian
n
o
is
e
at
d
if
f
er
en
t
SNR
v
alu
e
s
.
Fig
u
r
e
9
s
h
o
ws
th
e
cla
s
s
if
i
ca
tio
n
ac
cu
r
ac
y
o
f
m
o
to
r
b
ea
r
in
g
f
au
lts
f
o
r
f
iv
e
d
if
f
er
en
t
in
p
u
t
d
ata
s
izes
u
s
in
g
th
e
p
r
o
p
o
s
ed
MI
-
C
NN
m
o
d
el
u
s
in
g
two
-
p
h
ase
m
o
to
r
cu
r
r
en
t
s
ig
n
als
at
an
SNR
o
f
0
d
B
with
L
1
×
L2
=
40
×
4
0
,
6
0
×
6
0
,
8
0
×
8
0
,
an
d
1
0
0
×
1
0
0
i
n
Fig
u
r
es
9
(
a)
to
9
(
d
)
r
esp
ec
tiv
ely
.
T
h
e
ex
p
er
im
en
t
was
r
ep
ea
ted
with
o
t
h
er
SNR
v
alu
es
an
d
th
e
r
esu
lts
f
o
r
class
if
icatio
n
ac
cu
r
ac
y
an
d
th
e
tim
e
to
class
if
y
a
s
in
g
le
im
a
g
e
with
d
if
f
e
r
en
t
in
p
u
t
d
ata
s
i
ze
s
ar
e
p
r
esen
ted
in
T
ab
les 1
1
an
d
1
2
,
r
esp
ec
tiv
ely
.
4
.
5
.
F
a
ult
c
la
s
s
if
ica
t
io
n a
cc
ura
cy
co
m
pa
riso
n o
f
pro
po
s
ed
M
I
-
CNN
m
o
del a
nd
ba
s
ic
CNN
m
o
del
I
n
th
is
e
x
p
er
im
e
n
t,
we
u
s
ed
t
h
e
p
r
o
p
o
s
ed
MI
-
C
NN
m
o
d
el
w
ith
two
-
p
h
ase
m
o
to
r
c
u
r
r
en
t
s
ig
n
als
an
d
th
e
b
asic
C
NN
m
o
d
el
with
s
ig
n
als
f
r
o
m
p
h
ase
1
o
r
p
h
ase
2
.
T
h
e
p
ar
am
eter
s
f
o
r
th
e
C
NN
m
o
d
el
ar
e
d
etailed
in
T
ab
le
1
3
.
W
ith
an
in
p
u
t
i
m
ag
e
s
ize
o
f
8
0
×
8
0
,
ea
ch
b
ea
r
in
g
co
d
e
i
n
T
ab
le
3
h
as
2
5
.
6
0
0
.
0
0
0
/6
4
0
0
=4
0
0
0
im
ag
es,
with
3
2
0
0
im
ag
es
u
s
ed
f
o
r
tr
ai
n
in
g
an
d
8
0
0
im
ag
es
u
s
ed
f
o
r
test
in
g
.
E
x
p
e
r
i
m
e
n
ts
w
er
e
c
o
n
d
u
cte
d
wit
h
b
o
th
m
o
d
e
ls
u
s
i
n
g
d
at
a
w
ith
a
n
d
w
it
h
o
u
t
Ga
u
s
s
i
a
n
n
o
is
e
,
a
t
SNR
le
v
els o
f
2
0
,
1
5
,
1
0
,
5
,
0
,
-
5
,
a
n
d
-
1
0
d
B
an
d
c
o
m
p
a
r
ed
t
h
e
a
cc
u
r
ac
y
a
n
d
class
if
ica
ti
o
n
t
im
e
o
f
t
h
e
t
wo
m
o
d
els
.
Fig
u
r
e
9
.
Fau
lt c
lass
if
icatio
n
a
cc
u
r
ac
y
o
f
th
e
p
r
o
p
o
s
ed
MI
-
C
NN
m
o
d
el
with
d
if
f
e
r
en
t in
p
u
t
im
ag
e
s
izes,
ex
p
er
im
en
t w
ith
th
e
0
d
B
SN
R
s
ig
n
al
(
a)
L
1
×L
2
=
4
0
×4
0
,
(
b
)
(
b
)
L
1
×L
2
=
6
0
×6
0
,
(
c)
L
1
×L
2
=8
0
×8
0
,
an
d
(
d
)
L
1
×
L
2
=1
0
0
×1
0
0
T
ab
le
1
1
.
Fau
lt c
lass
if
icatio
n
ac
cu
r
ac
y
with
v
a
r
y
in
g
in
p
u
t im
ag
e
s
izes
I
n
p
u
t
i
m
a
g
e
s
i
z
e
F
a
u
l
t
c
l
a
ssi
f
i
c
a
t
i
o
n
a
c
c
u
r
a
c
y
(
%) w
i
t
h
n
o
i
s
e
-
a
d
d
i
n
g
s
i
g
n
a
l
a
t
d
i
f
f
e
r
e
n
t
S
N
R
v
a
l
u
e
s
2
0
d
B
1
5
d
B
1
0
d
B
5
d
B
0
d
B
-
5
d
B
-
1
0
d
B
40
40
8
1
.
2
2
7
2
.
3
8
6
6
.
2
2
6
0
.
9
4
5
4
.
3
0
4
8
.
4
0
4
4
.
8
0
60
60
7
2
.
0
5
6
6
.
8
2
5
9
.
1
9
5
2
.
4
8
4
8
.
6
0
4
6
.
3
7
4
2
.
8
5
80
80
8
1
.
9
6
8
0
.
2
0
7
4
.
4
0
6
4
.
6
7
5
5
.
9
0
4
9
.
7
9
4
7
.
8
2
1
0
0
1
0
0
7
9
.
8
0
7
6
.
4
4
6
2
.
0
8
5
6
.
3
4
5
1
.
5
0
4
7
.
1
8
4
4
.
6
8
T
ab
le
1
2
.
C
lass
if
icatio
n
tim
e
p
er
im
ag
e
with
v
ar
y
in
g
in
p
u
t i
m
ag
e
s
izes
I
n
p
u
t
i
m
a
g
e
s
i
z
e
N
u
mb
e
r
o
f
i
m
a
g
e
s
t
o
b
e
c
l
a
ss
i
f
i
e
d
To
t
a
l
w
e
i
g
h
t
(
w
i
t
h
F
=
1
1
1
1
,
K
=
1
)
Ti
me
t
o
c
l
a
s
si
f
y
o
n
e
i
m
a
g
e
(
ms)
40
40
9
6
0
0
4
3
6
4
3
1
.
2
3
±
0
.
1
60
60
4
2
6
7
4
5
6
4
3
2
.
5
0
±
0
.
2
80
80
2
4
0
0
5
2
0
4
3
3
.
4
0
±
0
.
3
1
0
0
1
0
0
1
5
3
6
5
6
4
4
3
6
.
3
9
±
0
.
6
T
ab
le
1
3
.
Par
a
m
eter
s
o
f
p
r
o
p
o
s
ed
MI
-
C
NN
m
o
d
el
P
a
r
a
me
t
e
r
V
a
l
u
e
P
a
r
a
me
t
e
r
V
a
l
u
e
I
n
p
u
t
i
m
a
g
e
s
i
z
e
80
80
L2
R
e
g
u
l
a
r
i
z
a
t
i
o
n
1e
-
10
M
i
n
i
b
a
t
c
h
s
i
z
e
1
2
8
Ep
o
c
h
10
I
n
i
t
i
a
l
Le
a
r
n
R
a
t
e
0
.
0
0
1
N
u
mb
e
r
o
f
k
e
r
n
e
l
s
/
K
e
r
n
e
l
si
z
e
i
n
c
o
n
v
o
l
u
t
i
o
n
a
l
l
a
y
e
r
s
5
/
3
3
Le
a
r
n
R
a
t
e
D
r
o
p
F
a
c
t
o
r
0
.
1
O
p
t
i
mi
z
e
r
A
d
a
m
V
a
l
i
d
a
t
i
o
n
P
a
t
i
e
n
c
e
5
N
u
mb
e
r
o
f
c
o
n
v
o
l
u
t
i
o
n
a
l
l
a
y
e
r
s
5
Evaluation Warning : The document was created with Spire.PDF for Python.