I
nte
rna
t
io
na
l J
o
urna
l o
f
P
o
wer
E
lect
ro
nics
a
nd
Driv
e
S
y
s
t
em
(
I
J
P
E
DS)
Vo
l.
1
7
,
No
.
1
,
Ma
r
ch
20
2
6
,
p
p
.
7
6
5
~
7
76
I
SS
N:
2
0
8
8
-
8
6
9
4
,
DOI
: 1
0
.
1
1
5
9
1
/ijp
ed
s
.
v
1
7
.
i
1
.
p
p
7
6
5
-
7
76
765
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
p
e
d
s
.
ia
esco
r
e.
co
m
Applica
tion o
f
ma
chine learni
ng
f
o
r
productio
n op
ti
miza
tion
a
nd predic
tive m
a
intenance in
an i
ro
n proces
sing
pl
a
nt
L
a
k
hd
a
ri
L
a
hcen,
M
o
ha
m
ed
H
a
bb
a
b,
Alha
chem
i Mo
ul
a
y
Abdella
h
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
i
c
a
l
En
g
i
n
e
e
r
i
n
g
,
La
b
o
r
a
t
o
r
y
o
f
C
o
n
t
r
o
l
A
n
a
l
y
si
s a
n
d
O
p
t
i
mi
z
a
t
i
o
n
o
f
El
e
c
t
r
o
-
E
n
e
r
g
e
t
i
c
S
y
st
e
ms (
C
A
O
S
EE),
F
a
c
u
l
t
y
o
f
T
e
c
h
n
o
l
o
g
y
,
Ta
h
r
i
M
o
h
a
m
med
U
n
i
v
e
r
si
t
y
,
B
e
c
h
a
r
,
A
l
g
e
r
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Au
g
1
1
,
2
0
2
5
R
ev
is
ed
J
an
5
,
2
0
2
6
Acc
ep
ted
J
an
2
3
,
2
0
2
6
T
h
e
m
o
d
e
r
n
m
e
t
a
l
l
u
r
g
i
c
a
l
i
n
d
u
s
tr
y
r
e
q
u
i
r
e
s
a
d
v
a
n
c
e
d
s
o
l
u
t
i
o
n
s
f
o
r
p
r
o
c
e
s
s
o
p
t
i
m
i
z
a
t
i
o
n
,
c
o
s
t
r
e
d
u
c
t
i
o
n
,
a
n
d
p
r
e
d
i
c
t
i
v
e
m
a
i
n
t
e
n
a
n
c
e
.
T
h
i
s
p
a
p
e
r
p
r
o
p
o
s
e
s
a
u
n
i
f
i
e
d
s
i
m
u
l
a
t
i
o
n
-
b
a
s
e
d
f
r
a
m
e
w
o
r
k
u
s
i
n
g
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
(
M
L
)
t
o
j
o
i
n
t
l
y
a
d
d
r
e
s
s
p
r
o
d
u
c
t
i
o
n
o
p
t
i
m
i
z
a
t
i
o
n
a
n
d
m
a
i
n
t
e
n
a
n
c
e
p
re
d
i
c
t
i
o
n
i
n
a
v
i
r
t
u
a
l
i
r
o
n
p
r
o
c
e
s
s
i
n
g
e
n
v
i
r
o
n
m
e
n
t
.
S
e
v
e
r
a
l
M
L
m
o
d
e
l
s
,
i
n
c
l
u
d
i
n
g
r
a
n
d
o
m
f
o
r
e
s
t
(R
F
)
,
e
x
t
r
e
m
e
g
r
a
d
ie
n
t
b
o
o
s
t
i
n
g
(
XG
B
o
o
s
t
)
,
l
i
g
h
t
g
r
a
d
i
e
n
t
b
o
o
s
t
i
n
g
m
a
c
h
i
n
e
(
L
i
g
h
t
G
B
M
)
,
s
u
p
p
o
r
t
v
e
c
t
o
r
m
a
c
h
i
n
e
(
S
V
M
)
,
a
n
d
k
-
n
e
a
r
e
s
t
n
e
i
g
h
b
o
r
s
(
k
-
N
N
)
,
w
e
r
e
e
v
a
l
u
a
t
e
d
o
n
s
y
n
t
h
e
t
i
c
d
a
t
a
s
e
t
s
re
p
r
e
se
n
t
i
n
g
p
r
o
d
u
c
t
i
o
n
,
m
a
i
n
t
e
n
a
n
c
e
,
a
n
d
t
r
a
n
s
p
o
r
t
p
r
o
c
e
s
s
e
s
.
A
r
e
p
r
o
d
u
c
i
b
l
e
m
e
t
h
o
d
o
l
o
g
y
w
a
s
a
d
o
p
t
e
d
,
i
n
c
l
u
d
i
n
g
p
r
e
p
r
o
c
e
s
s
i
n
g
,
t
im
e
-
a
w
a
r
e
d
a
ta
s
p
l
i
t
t
i
n
g
,
a
n
d
c
r
o
s
s
-
v
a
l
i
d
a
t
i
o
n
to
p
r
e
v
e
n
t
i
n
f
o
r
m
a
t
i
o
n
l
e
a
k
a
g
e
.
M
o
d
e
l
p
e
r
f
o
r
m
a
n
c
e
wa
s
a
s
se
s
se
d
u
s
i
n
g
F
1
-
s
c
o
r
e
,
a
re
a
u
n
d
e
r
t
h
e
r
e
c
e
i
v
e
r
o
p
e
r
a
t
i
n
g
c
h
a
r
a
c
t
e
r
i
s
t
ic
c
u
r
v
e
(
A
U
C
)
,
a
n
d
r
e
g
r
e
s
s
i
o
n
m
e
t
r
ic
s
.
T
re
e
-
b
a
s
e
d
m
o
d
e
l
s
a
c
h
i
e
v
e
d
n
e
a
r
-
p
e
r
fe
c
t
c
l
a
s
s
i
f
ic
a
t
i
o
n
p
e
r
f
o
r
m
a
n
c
e
(
A
U
C
≈
1
,
p
r
e
c
i
s
i
o
n
a
n
d
r
e
c
a
l
l
>
0
.
9
9
)
,
w
h
i
l
e
l
i
g
h
t
g
r
a
d
i
e
n
t
b
o
o
s
t
i
n
g
m
a
c
h
i
n
e
(
L
i
g
h
t
G
B
M
)
a
n
d
C
a
t
B
o
o
s
t
p
r
o
v
i
d
e
d
t
h
e
b
e
s
t
r
e
g
r
e
s
s
i
o
n
a
c
c
u
r
a
c
y
.
F
e
a
t
u
r
e
i
m
p
o
r
t
a
n
c
e
a
n
a
l
y
s
i
s
u
s
i
n
g
S
H
a
p
l
e
y
A
d
d
i
t
i
v
e
e
x
P
l
a
n
a
t
i
o
n
s
(
S
H
A
P
)
i
d
e
n
t
i
f
i
e
d
v
i
b
r
a
t
i
o
n
a
n
d
t
e
m
p
e
r
a
t
u
r
e
a
s
k
e
y
m
a
i
n
t
e
n
a
n
c
e
i
n
d
i
c
a
t
o
r
s
.
A
l
t
h
o
u
g
h
b
a
s
e
d
o
n
s
i
m
u
l
a
t
i
o
n
,
t
h
e
f
r
a
m
e
w
o
r
k
i
s
d
e
s
ig
n
e
d
f
o
r
i
n
t
e
g
r
a
t
i
o
n
w
i
t
h
s
u
p
e
r
v
i
s
o
r
y
c
o
n
t
r
o
l
a
n
d
d
a
t
a
a
c
q
u
i
s
i
t
i
o
n
(
S
C
A
DA
)
a
n
d
t
h
e
I
n
d
u
s
t
r
i
a
l
I
n
t
e
r
n
e
t
o
f
T
h
i
n
g
s
(
I
I
o
T
)
,
s
u
p
p
o
r
t
i
n
g
r
e
a
l
-
t
i
m
e
i
n
d
u
s
t
r
i
a
l
d
e
p
l
o
y
m
e
n
t
a
n
d
a
l
i
g
n
m
e
n
t
w
i
t
h
o
p
e
r
a
t
i
o
n
a
l
k
e
y
p
e
r
f
o
r
m
a
n
c
e
i
n
d
i
c
a
t
o
r
s
.
K
ey
w
o
r
d
s
:
I
r
o
n
p
r
o
ce
s
s
in
g
p
la
n
t
L
ig
h
tGB
M
Ma
ch
in
e
lear
n
in
g
Pre
d
ictiv
e
m
ain
ten
an
ce
Pro
d
u
ctio
n
o
p
tim
izatio
n
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
:
L
ak
h
d
ar
i L
a
h
ce
n
Dep
ar
tm
en
t o
f
E
lectr
ical
E
n
g
i
n
ee
r
in
g
L
ab
o
r
ato
r
y
o
f
C
o
n
tr
o
l A
n
al
y
s
is
an
d
Op
tim
izatio
n
o
f
E
lectr
o
-
E
n
er
g
etic
Sy
s
tem
s
(
C
AOSE
E
)
Facu
lty
o
f
T
ec
h
n
o
lo
g
y
,
T
a
h
r
i
Mo
h
am
m
ed
U
n
iv
er
s
ity
B
ec
h
ar
,
Alg
er
ia
E
m
ail:
lak
h
d
ar
i.la
h
ce
n
@
u
n
iv
-
b
ec
h
ar
.
d
z
1.
I
NT
RO
D
UCT
I
O
N
T
h
e
m
etallu
r
g
ical
in
d
u
s
tr
y
i
s
a
co
r
n
er
s
to
n
e
o
f
ec
o
n
o
m
i
c
d
ev
elo
p
m
en
t,
s
u
p
p
ly
in
g
cr
itical
r
a
w
m
ater
ials
s
u
ch
as
ir
o
n
an
d
s
teel
f
o
r
s
ec
to
r
s
in
clu
d
in
g
co
n
s
tr
u
ctio
n
,
au
to
m
o
tiv
e
m
a
n
u
f
ac
tu
r
in
g
,
in
f
r
astru
ctu
r
e
,
an
d
en
er
g
y
.
Am
o
n
g
th
e
r
aw
m
ater
ials
,
ir
o
n
o
r
e
h
o
l
d
s
a
p
i
v
o
tal
r
o
le
as
it
is
th
e
p
r
im
ar
y
f
ee
d
s
to
ck
f
o
r
s
teel
p
r
o
d
u
ctio
n
.
T
h
e
d
em
an
d
f
o
r
h
ig
h
-
q
u
ality
ir
o
n
o
r
e
is
s
te
ad
ily
in
cr
ea
s
in
g
,
d
r
i
v
en
b
y
g
lo
b
al
u
r
b
an
izatio
n
,
in
d
u
s
tr
ializatio
n
,
an
d
th
e
tr
an
s
itio
n
to
m
o
d
er
n
,
e
n
er
g
y
-
ef
f
icie
n
t in
f
r
astru
ctu
r
e
[
1
]
,
[
2
]
.
I
n
Al
g
e
r
ia
,
t
h
e
G
h
ar
D
je
b
il
et
i
r
o
n
o
r
e
d
e
p
o
s
it
,
l
o
ca
te
d
i
n
t
h
e
T
i
n
d
o
u
f
P
r
o
v
i
n
ce
in
th
e
s
o
u
t
h
west
o
f
t
h
e
co
u
n
t
r
y
,
r
e
p
r
es
en
ts
o
n
e
o
f
t
h
e
la
r
g
est
u
n
ta
p
p
e
d
ir
o
n
r
es
er
v
e
s
in
t
h
e
w
o
r
l
d
,
wit
h
esti
m
a
te
d
d
e
p
o
s
its
e
x
c
ee
d
i
n
g
3
.
5
b
il
li
o
n
t
o
n
n
es
.
T
h
is
s
t
r
a
te
g
ic
r
eso
u
r
c
e
h
as
t
h
e
p
o
te
n
ti
al
t
o
t
r
a
n
s
f
o
r
m
Al
g
e
r
i
a
i
n
t
o
a
k
e
y
p
la
y
er
i
n
t
h
e
g
l
o
b
a
l
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
6
9
4
I
n
t J Po
w
E
lec
&
Dr
i Sy
s
t
,
Vo
l.
1
7
,
No
.
1
,
Ma
r
c
h
20
2
6
:
765
-
776
766
s
teel
in
d
u
s
t
r
y
.
T
h
e
p
la
n
n
e
d
d
e
v
el
o
p
m
e
n
t
o
f
e
x
tr
ac
ti
o
n
,
p
r
o
ce
s
s
in
g
,
an
d
tr
an
s
p
o
r
t
i
n
f
r
astr
u
ct
u
r
e
ass
o
c
iat
e
d
wit
h
Gh
a
r
Dj
eb
ile
t w
ill
n
o
t
o
n
l
y
m
e
et
d
o
m
esti
c
s
tee
l
d
e
m
a
n
d
b
u
t
a
ls
o
e
n
a
b
l
e
s
i
g
n
i
f
i
ca
n
t
e
x
p
o
r
t
o
p
p
o
r
t
u
n
iti
es
[
1
]
.
Ho
w
ev
er
,
t
h
e
ex
p
l
o
i
tat
io
n
o
f
s
u
c
h
a
l
ar
g
e
-
s
ca
l
e
m
i
n
i
n
g
a
n
d
p
r
o
ce
s
s
in
g
p
r
o
j
ec
t
p
r
es
en
t
s
c
o
m
p
le
x
tec
h
n
ic
al
a
n
d
o
p
e
r
ati
o
n
a
l
c
h
a
lle
n
g
es.
T
h
es
e
i
n
cl
u
d
e
o
p
ti
m
i
zin
g
p
r
o
d
u
c
ti
o
n
p
a
r
am
ete
r
s
,
en
s
u
r
i
n
g
c
o
n
s
is
te
n
t
p
r
o
d
u
ct
q
u
ali
ty
,
a
n
d
i
m
p
le
m
en
t
in
g
p
r
ed
ict
iv
e
m
a
in
te
n
a
n
c
e
s
t
r
a
te
g
ies
t
o
m
i
n
i
m
i
ze
e
q
u
ip
m
e
n
t
d
o
w
n
t
im
e.
T
r
ad
iti
o
n
al
in
d
u
s
t
r
ia
l
c
o
n
t
r
o
l
s
y
s
te
m
s
o
f
te
n
r
el
y
o
n
s
tat
ic
p
r
o
ce
s
s
r
u
les
a
n
d
p
e
r
i
o
d
ic
m
a
n
u
al
i
n
s
p
e
cti
o
n
s
,
w
h
i
c
h
m
a
y
b
e
i
n
s
u
f
f
i
ci
en
t i
n
a
h
i
g
h
-
t
h
r
o
u
g
h
p
u
t
,
g
e
o
g
r
a
p
h
ic
all
y
r
em
o
te
i
n
d
u
s
tr
ial
e
n
v
i
r
o
n
m
e
n
t
[
3
]
.
T
h
e
ad
v
e
n
t
o
f
I
n
d
u
s
tr
y
4
.
0
te
ch
n
o
lo
g
ies
,
in
teg
r
atin
g
s
en
s
o
r
s
,
r
ea
l
-
tim
e
d
ata
ac
q
u
is
itio
n
,
ad
v
an
ce
d
an
aly
tics
,
an
d
ar
tific
ial
in
telli
g
en
ce
(
AI
)
,
o
f
f
e
r
s
u
n
p
r
ec
ed
e
n
ted
o
p
p
o
r
tu
n
ities
to
m
o
d
er
n
iz
e
th
e
ir
o
n
an
d
s
teel
v
alu
e
ch
ain
.
W
ith
in
th
is
tec
h
n
o
lo
g
ical
ec
o
s
y
s
tem
,
m
ac
h
i
n
e
lear
n
in
g
(
ML
)
h
as
em
er
g
ed
as
a
p
o
wer
f
u
l
ap
p
r
o
ac
h
f
o
r
e
x
tr
ac
tin
g
ac
tio
n
ab
le
in
s
ig
h
ts
f
r
o
m
lar
g
e
a
n
d
c
o
m
p
lex
d
atasets
.
ML
m
o
d
els
ca
n
id
en
tify
s
u
b
tle
p
atter
n
s
,
ad
ap
t
to
ev
o
lv
in
g
o
p
er
atio
n
al
co
n
d
itio
n
s
,
an
d
s
u
p
p
o
r
t
d
ata
-
d
r
iv
en
d
ec
is
io
n
-
m
ak
in
g
in
b
o
th
p
r
o
d
u
ctio
n
an
d
m
ain
ten
a
n
ce
d
o
m
ain
s
[
2
]
‒
[
4
]
.
I
n
th
is
co
n
tex
t,
th
e
o
b
jectiv
e
o
f
th
is
r
esear
ch
is
to
in
v
esti
g
ate
th
e
ap
p
licatio
n
o
f
s
u
p
e
r
v
is
ed
ML
alg
o
r
ith
m
s
to
o
p
tim
ize
p
r
o
d
u
c
tio
n
q
u
ality
a
n
d
p
r
ed
ict
eq
u
i
p
m
en
t
f
ailu
r
es
in
an
ir
o
n
p
r
o
ce
s
s
in
g
p
lan
t
in
s
p
ir
ed
b
y
th
e
o
p
er
atio
n
al
n
ee
d
s
o
f
t
h
e
Gh
ar
Djeb
ilet
p
r
o
ject.
Giv
en
th
e
ab
s
en
ce
o
f
r
ea
l
o
p
e
r
atio
n
al
d
ata
f
r
o
m
th
e
f
u
tu
r
e
f
ac
ilit
y
,
s
y
n
t
h
etic
d
ata
s
ets
h
av
e
b
ee
n
g
en
er
ate
d
to
s
im
u
late
r
ea
lis
tic
in
d
u
s
tr
ial
s
ce
n
ar
io
s
,
in
clu
d
in
g
p
r
o
d
u
ctio
n
,
m
ai
n
ten
an
ce
,
an
d
tr
an
s
p
o
r
t o
p
er
atio
n
s
[
4
]
‒
[
6
]
.
T
h
e
co
n
tr
ib
u
ti
o
n
s
o
f
th
is
s
tu
d
y
ar
e:
i)
Dev
elo
p
m
en
t
o
f
s
im
u
latio
n
-
b
ased
d
atasets
r
ep
r
esen
tin
g
k
ey
o
p
er
atio
n
al
p
r
o
ce
s
s
es
o
f
a
n
ir
o
n
p
r
o
ce
s
s
in
g
p
lan
t
;
ii)
C
o
m
p
ar
ativ
e
ev
al
u
atio
n
o
f
f
i
v
e
wid
ely
u
s
ed
ML
alg
o
r
ith
m
s
,
r
a
n
d
o
m
f
o
r
est
(
R
F)
,
ex
tr
em
e
g
r
a
d
ien
t
b
o
o
s
tin
g
(
XGBo
o
s
t)
,
lig
h
t
g
r
a
d
ien
t
b
o
o
s
tin
g
m
ac
h
in
e
(
L
ig
h
tGB
M)
,
ca
teg
o
r
ical
b
o
o
s
tin
g
(
C
atB
o
o
s
t)
,
n
eu
r
al
n
etwo
r
k
(
NN)
,
an
d
R
F+NN
f
o
r
class
if
icatio
n
(
p
r
ed
ictiv
e
m
ain
ten
a
n
ce
)
a
n
d
r
eg
r
ess
io
n
(
p
r
o
d
u
ctio
n
q
u
alit
y
)
;
iii)
I
n
teg
r
atio
n
o
f
i
n
ter
p
r
e
tab
ilit
y
tech
n
iq
u
es
(
f
ea
tu
r
e
im
p
o
r
tan
ce
,
SHap
ley
Ad
d
itiv
e
ex
Plan
atio
n
s
(
SHAP
)
)
to
en
h
an
ce
m
o
d
el
tr
an
s
p
a
r
en
cy
an
d
s
u
p
p
o
r
t
in
d
u
s
tr
ial
ad
o
p
tio
n
;
an
d
iv
)
F
o
r
m
u
latio
n
o
f
a
m
eth
o
d
o
lo
g
ic
al
f
r
am
ewo
r
k
ad
ap
tab
le
to
r
ea
l
in
d
u
s
tr
ial
d
ataset
s
f
o
r
f
u
tu
r
e
d
ep
lo
y
m
en
t in
th
e
Gh
ar
Djeb
ilet ir
o
n
p
r
o
ce
s
s
in
g
c
h
ain
.
I
n
p
ar
allel
with
th
e
n
atio
n
al
d
ev
elo
p
m
e
n
t
o
f
t
h
e
Gh
a
r
Dj
eb
ilet
m
in
in
g
p
r
o
ject,
a
d
e
d
i
ca
ted
ir
o
n
-
p
r
o
ce
s
s
in
g
p
lan
t
h
as
b
ee
n
esta
b
lis
h
ed
in
th
e
B
ec
h
ar
r
eg
io
n
o
f
s
o
u
th
wester
n
Alg
er
ia.
T
h
is
i
n
d
u
s
tr
ial
f
ac
ilit
y
is
ex
p
ec
ted
t
o
b
ec
o
m
e
a
k
e
y
c
o
m
p
o
n
e
n
t
in
tr
an
s
f
o
r
m
in
g
r
aw
ex
tr
ac
ted
o
r
e
in
to
s
em
i
-
f
in
is
h
ed
o
r
f
in
is
h
ed
ir
o
n
p
r
o
d
u
cts,
wh
ile
also
s
er
v
in
g
as
a
tech
n
o
lo
g
ical
p
ilo
t
p
la
tf
o
r
m
f
o
r
test
in
g
an
d
d
e
p
lo
y
in
g
m
o
d
er
n
d
ig
ital
s
o
lu
tio
n
s
in
m
etallu
r
g
y
.
I
n
th
i
s
s
tu
d
y
,
th
e
B
ec
h
ar
p
lan
t
is
ad
o
p
ted
as
th
e
r
ef
er
en
ce
in
d
u
s
tr
ial
en
v
ir
o
n
m
e
n
t.
All
s
im
u
latio
n
s
an
d
p
r
ed
ictiv
e
m
o
d
els
h
av
e
b
ee
n
d
esig
n
ed
i
n
alig
n
m
en
t
with
th
e
an
ticip
ate
d
co
n
f
ig
u
r
atio
n
a
n
d
wo
r
k
f
lo
ws
o
f
th
is
f
ac
ilit
y
.
I
ts
p
r
o
x
im
ity
to
T
ah
r
i
Mo
h
am
m
ed
Un
i
v
er
s
ity
o
f
B
ec
h
ar
an
d
th
e
ac
tiv
e
in
v
o
lv
em
e
n
t
o
f
lo
ca
l
ac
ad
e
m
ic
r
esear
ch
in
th
e
f
ield
s
o
f
au
t
o
m
atio
n
a
n
d
in
d
u
s
tr
ial
in
f
o
r
m
atics
o
f
f
er
a
u
n
iq
u
e
o
p
p
o
r
tu
n
ity
to
b
r
id
g
e
th
e
g
a
p
b
etwe
en
ad
v
an
ce
d
ML
r
esear
ch
an
d
r
ea
l
-
w
o
r
ld
d
e
p
lo
y
m
en
t
in
a
s
tr
ateg
ic
n
atio
n
al
in
d
u
s
tr
ial
p
r
o
ject
[
7
]
‒
[
9
]
.
W
e
ad
d
r
ess
q
u
ality
s
tab
ilizat
io
n
,
th
r
o
u
g
h
p
u
t,
an
d
d
o
wn
ti
m
e
r
ed
u
ctio
n
i
n
ir
o
n
p
r
o
ce
s
s
in
g
,
wh
e
r
e
co
n
v
en
tio
n
al
r
u
le
-
b
ased
c
o
n
tr
o
l
an
d
p
er
i
o
d
ic
in
s
p
ec
tio
n
s
f
a
lter
u
n
d
e
r
m
u
ltiv
ar
iate
n
o
n
-
li
n
ea
r
ities
,
d
r
if
t,
a
n
d
n
o
is
e.
Un
lik
e
p
r
i
o
r
wo
r
k
th
at
is
o
lates
a
s
in
g
le
s
u
b
p
r
o
b
lem
an
d
o
m
its
r
ig
o
r
o
u
s
v
alid
atio
n
an
d
d
e
p
lo
y
a
b
ilit
y
,
we
p
r
o
p
o
s
e
a
u
n
if
ied
ML
f
r
a
m
ewo
r
k
o
v
e
r
p
r
o
d
u
ctio
n
,
m
a
in
ten
an
ce
,
an
d
tr
a
n
s
p
o
r
t
u
s
in
g
s
im
u
latio
n
-
b
ased
d
atasets
.
W
e
b
en
ch
m
ar
k
R
F,
XGBo
o
s
t,
L
ig
h
tGB
M,
C
at
B
o
o
s
t,
ML
P,
SVM,
k
-
NN,
a
n
d
R
F+NN,
with
tim
e
-
awa
r
e
s
p
lits
,
cr
o
s
s
-
v
alid
atio
n
,
laten
cy
-
awa
r
e
s
elec
tio
n
,
an
d
SHAP
-
b
ased
in
ter
p
r
etatio
n
.
R
esu
lts
(
ar
ea
u
n
d
er
th
e
r
ec
eiv
er
o
p
e
r
atin
g
c
h
ar
ac
ter
is
tic
(
R
OC
)
cu
r
v
e
(
AUC
)
≈
1
,
p
r
ec
is
io
n
/r
ec
all
>
0
.
9
9
f
o
r
class
if
icatio
n
;
co
ef
f
icien
t
o
f
d
eter
m
in
atio
n
(
R
2
)
>
0
.
5
9
f
o
r
r
eg
r
ess
io
n
with
b
o
o
s
ter
s
)
ar
e
lin
k
ed
to
k
ey
p
er
f
o
r
m
a
n
ce
in
d
icato
r
s
(
KPI
)
im
p
ac
ts
(
d
o
wn
tim
e,
m
ain
ten
an
ce
,
th
r
o
u
g
h
p
u
t
in
Alg
er
ian
Din
ar
(
DZ
D
)
)
an
d
a
r
ea
l
-
tim
e
s
u
p
er
v
is
o
r
y
co
n
tr
o
l
an
d
d
ata
ac
q
u
is
itio
n
(
SC
ADA
)
/
in
d
u
s
tr
ia
l
in
ter
n
et
o
f
th
in
g
s
(
I
I
o
T
)
d
ep
l
o
y
m
en
t
p
ath
(
o
p
en
p
latf
o
r
m
co
m
m
u
n
icatio
n
s
u
n
if
ied
ar
ch
itectu
r
e
(
OPC
UA
)
/
m
ess
ag
e
q
u
eu
in
g
telem
etr
y
tr
an
s
p
o
r
t
(
MQ
T
T
)
,
o
p
en
n
eu
r
al
n
etwo
r
k
ex
c
h
an
g
e
(
ONNX
)
,
e
d
g
e
i
n
f
er
en
ce
)
.
T
h
is
r
esear
ch
in
v
esti
g
ates
th
e
ap
p
licatio
n
o
f
s
u
p
er
v
is
ed
m
ac
h
i
n
e
lear
n
in
g
alg
o
r
ith
m
s
to
o
p
tim
ize
p
r
o
d
u
ctio
n
q
u
ality
a
n
d
p
r
ed
ict
eq
u
i
p
m
en
t
f
ailu
r
es
in
an
ir
o
n
p
r
o
ce
s
s
in
g
p
lan
t.
T
h
e
o
p
e
r
atio
n
al
n
ee
d
s
o
f
th
e
Gh
a
r
Djeb
ilet p
r
o
ject
in
s
p
ir
e
th
e
s
tu
d
y
.
T
h
e
in
d
u
s
tr
ial
co
n
tex
t
o
f
th
i
s
s
tu
d
y
is
illu
s
tr
ated
in
F
ig
u
r
e
1
,
wh
ich
s
h
o
ws
th
e
o
p
en
-
p
it
q
u
ar
r
y
s
tr
u
ctu
r
e
o
f
th
e
Gar
a
Djeb
ilet
m
in
in
g
s
ite.
T
h
e
lar
g
e
-
s
ca
le
ex
tr
ac
tio
n
en
v
ir
o
n
m
en
t
in
v
o
lv
e
s
h
ea
v
y
eq
u
ip
m
en
t,
co
n
tin
u
o
u
s
m
ater
ial
tr
an
s
p
o
r
t,
an
d
d
y
n
am
ic
o
p
er
atin
g
c
o
n
d
itio
n
s
,
h
ig
h
lig
h
tin
g
th
e
im
p
o
r
tan
ce
o
f
p
r
ed
ictiv
e
m
ain
ten
an
ce
an
d
i
n
tellig
en
t
p
r
o
d
u
ctio
n
m
a
n
ag
em
e
n
t.
Fig
u
r
e
2
p
r
esen
ts
a
r
ep
r
esen
tativ
e
v
i
ew
o
f
an
in
d
u
s
tr
ial
ir
o
n
-
p
r
o
ce
s
s
in
g
f
ac
ilit
y
.
T
h
e
co
m
p
le
x
a
r
r
an
g
e
m
en
t
o
f
p
r
o
ce
s
s
in
g
u
n
its
an
d
m
o
n
ito
r
in
g
in
f
r
astru
ctu
r
es
g
en
er
ates
h
eter
o
g
en
e
o
u
s
o
p
er
atio
n
al
d
ata
s
tr
ea
m
s
th
at
c
an
b
e
ef
f
ec
tiv
ely
e
x
p
lo
ited
u
s
in
g
m
ac
h
in
e
lear
n
i
n
g
tec
h
n
iq
u
es to
im
p
r
o
v
e
s
y
s
tem
r
eliab
ilit
y
,
ef
f
icien
cy
,
an
d
d
ec
i
s
io
n
-
m
ak
in
g
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Po
w
E
lec
&
Dr
i Sy
s
t
I
SS
N:
2088
-
8
6
9
4
A
p
p
lica
tio
n
o
f m
a
ch
in
e
le
a
r
n
in
g
fo
r
p
r
o
d
u
ctio
n
o
p
timiz
a
tio
n
a
n
d
…
(
La
kh
d
a
r
i La
h
ce
n
)
767
Fig
u
r
e
1
.
I
n
ter
io
r
v
iew
o
f
t
h
e
Gar
a
Djeb
ilet
q
u
ar
r
y
s
tr
u
ctu
r
es illu
s
tr
ates th
e
o
p
en
-
p
it d
ep
o
s
it a
n
d
t
h
e
m
in
in
g
co
n
tex
t
Fig
u
r
e
2
.
View
o
f
r
o
b
u
s
t in
d
u
s
tr
ial
u
s
ef
u
l a
s
a
s
ch
em
atic
illu
s
tr
atio
n
2.
RE
A
L
-
T
I
M
E
DE
P
L
O
Y
M
E
NT
F
E
A
SI
B
I
L
I
T
Y
SUP
E
RVI
SO
RY
CO
N
T
RO
L
AND
DA
T
A
ACQ
UIS
I
T
I
O
N
(
SCADA)
Alth
o
u
g
h
all
e
x
p
er
im
e
n
ts
ar
e
co
n
d
u
cted
in
s
im
u
latio
n
,
th
e
ap
p
r
o
ac
h
is
co
m
p
atib
le
with
r
ea
l
-
tim
e
in
d
u
s
tr
ial
d
ep
lo
y
m
en
t w
ith
m
i
n
o
r
ad
a
p
tatio
n
[
1
0
]
,
[
1
1
]
.
a)
I
n
teg
r
atio
n
SC
ADA
−
T
ar
g
et
ar
ch
itectu
r
e
:
PLC/
SC
A
DA
⟶
ed
g
e
g
atew
ay
(
OPC
UA
o
r
Mo
d
b
u
s
/TCP
clien
t,
MQ
T
T
b
r
o
k
er
)
⟶
in
f
er
en
ce
m
icr
o
s
er
v
ice
(
R
F,
L
ig
h
tGB
M,
XGBo
o
s
t,
ML
P,
o
r
R
F+NN)
⟶
alar
m
f
ee
d
b
a
ck
t
o
SC
ADA
(
O
PC
UA
wr
ite)
an
d
d
ata
h
is
to
r
izatio
n
(
I
n
f
lu
x
DB
/T
im
escale/PI
)
.
−
Mo
d
el
p
ac
k
ag
in
g
:
ONNX
ex
p
o
r
t
(
On
n
x
r
u
n
tim
e
o
n
C
PU)
o
r
n
ativ
e
f
o
r
m
at
(
XGBo
o
s
t/Lig
h
tGB
M)
.
T
h
is
en
s
u
r
es p
o
r
tab
ilit
y
(
C
++
/
Py
th
o
n
)
a
n
d
s
tar
tu
p
tim
e
<
1
s
.
b)
L
aten
cy
b
u
d
g
et
an
d
s
am
p
lin
g
r
ate
E
n
d
-
to
-
e
n
d
laten
c
y
s
h
o
u
l
d
r
e
m
ain
b
elo
w
a
f
r
ac
tio
n
o
f
th
e
s
am
p
lin
g
p
e
r
io
d
:
T
t
o
t
al
=
T
acq
+
T
n
et
+
T
pre
+
T
i
n
f
+
T
p
ub
,
a
im
for
T
t
o
t
al
≤
0
.
3
×
(
s
a
mpl
in
g
pe
r
i
od
)
(
1
)
wh
er
e
T
_
ac
q
=
d
ata
ac
q
u
is
itio
n
tim
e,
T
_
n
et
=
n
etwo
r
k
co
m
m
u
n
icatio
n
tim
e,
T
_
p
r
e
=
d
ata
p
r
ep
r
o
ce
s
s
in
g
tim
e,
T
_
in
f
=
m
o
d
el
in
f
er
en
ce
tim
e,
T
_
p
u
b
=
p
u
b
lis
h
in
g
tim
e
(
wr
itin
g
b
ac
k
to
SC
ADA
o
r
to
d
atab
ases
)
,
an
d
T
_
to
tal
=
en
d
-
to
-
en
d
laten
cy
o
f
th
e
p
r
o
ce
s
s
in
g
ch
ain
.
E
x
a
m
p
les:
1
0
Hz
(
1
0
0
m
s
)
⇒
tar
g
et
≤
3
0
m
s
;
1
0
0
Hz
(
1
0
m
s
)
⇒
tar
g
et
≤
3
m
s
.
Use
in
cr
em
en
tal/s
tr
ea
m
in
g
f
ea
t
u
r
e
u
p
d
ates
wh
en
win
d
o
win
g
(
e.
g
.
,
1
s
win
d
o
w,
1
0
0
m
s
h
o
p
)
to
av
o
id
ad
d
e
d
d
elay
.
Har
d
war
e
co
n
s
tr
ain
ts
(
ed
g
e
)
:
−
C
PU x
8
6
/AR
M
I
P
C
: su
f
f
icien
t f
o
r
co
m
p
ac
t RF
/GB
DT
/ML
P
at
1
0
–
1
0
0
Hz.
−
L
ig
h
t G
PU/NPU (
J
et
s
o
n
,
I
n
tel
NPU)
: u
s
ef
u
l f
o
r
lar
g
er
ML
P/
C
NN
o
r
m
u
lti
-
s
tr
ea
m
b
atch
in
g
.
−
Me
m
o
r
y
b
u
d
g
et:
k
ee
p
s
er
v
ic
e
<
2
0
0
MB
(
m
o
d
el
+
r
u
n
tim
e
+
b
u
f
f
er
s
)
.
k
-
NN
is
R
AM
-
h
ea
v
y
;
p
r
ef
er
R
F/G
B
DT
/ML
P o
r
R
F
+N
N.
−
I
n
ter
f
ac
es: OPC
UA/M
o
d
b
u
s
/TCP
f
o
r
ac
q
u
is
itio
n
; M
QT
T
f
o
r
p
u
b
/s
u
b
; secu
r
e
with
T
L
S a
n
d
R
B
AC
.
c)
Or
d
er
-
of
-
m
ag
n
itu
d
e
in
f
er
en
ce
laten
cy
(
m
o
d
er
n
e
d
g
e
C
PU)
−
R
an
d
o
m
f
o
r
est (
1
0
0
–
5
0
0
tr
ee
s
,
d
ep
th
≤
1
0
)
: 0
.
1
–
2
m
s
/s
am
p
le
−
L
ig
h
tGB
M/XG
B
o
o
s
t
(
d
ep
th
≤
8
)
: 0
.
2
–
3
m
s
−
L
in
ea
r
SVM:
<
0
.
1
–
0
.
5
m
s
; RB
F SVM
: 1
–
1
0
m
s
(
d
ep
en
d
s
o
n
#
SV)
−
ML
P (
2
–
3
lay
e
r
s
,
6
4
–
1
2
8
)
: 0
.
2
–
2
m
s
−
R
F+NN
(
s
tack
in
g
)
: 1
–
4
m
s
d)
Dep
lo
y
m
en
t
p
atter
n
s
−
Har
d
r
ea
l
-
tim
e
(
≤
1
0
m
s
in
PLC):
av
o
id
co
m
p
lex
ML
; u
s
e
r
u
les/
lin
ea
r
/s
h
allo
w
tr
ee
s
.
−
So
f
t r
ea
l
-
tim
e
(
1
0
–
2
0
0
m
s
o
n
ed
g
e
g
atew
ay
)
: c
o
m
p
ac
t RF
/GB
DT
/ML
P;
R
F+NN
i
s
f
ea
s
ib
l
e.
−
Of
f
lin
e
m
o
n
ito
r
in
g
(
o
n
-
p
r
em
/c
lo
u
d
)
: r
etr
ai
n
in
g
,
r
ec
alib
r
atio
n
,
co
n
tr
o
lled
r
o
llo
u
ts
.
On
e
-
s
en
ten
ce
tak
ea
way
:
Pac
k
ag
ed
in
ONNX
an
d
e
x
ec
u
t
ed
o
n
an
e
d
g
e
g
atew
ay
v
ia
OPC
UA/M
QT
T
,
co
m
p
ac
t
R
F/GB
DT
/ML
P
(
o
r
R
F+NN)
ty
p
ically
ac
h
iev
e
≤
1
–
5
m
s
p
er
s
am
p
le,
s
u
f
f
icien
t
f
o
r
1
0
–
1
0
0
H
z
o
p
er
atio
n
i
f
p
r
e
p
r
o
ce
s
s
in
g
an
d
n
etwo
r
k
in
g
s
tay
with
in
b
u
d
g
e
t.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
6
9
4
I
n
t J Po
w
E
lec
&
Dr
i Sy
s
t
,
Vo
l.
1
7
,
No
.
1
,
Ma
r
c
h
20
2
6
:
765
-
776
768
3.
E
CO
NO
M
I
C
AND
O
P
E
RA
T
I
O
NA
L
I
M
P
AC
T
I
ND
I
CA
T
O
RS
Pu
r
p
o
s
e
:
C
o
n
v
er
t
m
o
d
el
p
er
f
o
r
m
an
ce
(
e.
g
.
,
R
F,
L
ig
h
tGB
M,
R
F+NN)
in
to
DA
g
ain
s
f
o
r
f
ac
to
r
y
KPI
s
[
1
2
]
,
[
1
3
]
.
T
h
e
k
e
y
ec
o
n
o
m
ic
an
d
o
p
er
atio
n
al
in
d
icato
r
s
ar
e
d
ef
in
ed
as f
o
llo
ws:
-
Av
ailab
ilit
y
=
−
(
2
)
-
R
eliab
ilit
y
an
d
m
ain
tai
n
ab
ilit
y
ar
e
ev
alu
ated
u
s
in
g
m
ea
n
tim
e
b
etwe
en
f
ailu
r
es
(
MT
B
F)
an
d
m
ea
n
tim
e
to
r
ep
air
(
MT
T
R
)
,
Ov
er
all
eq
u
ip
m
e
n
t e
f
f
ec
tiv
e
n
e
s
s
(
OE
E
)
=
×
×
(
3
)
-
Mo
n
etiza
tio
n
Alg
er
ian
Din
a
r
(
DA)
-
Do
wn
tim
e
s
av
ed
DA
≈
(
−
)
+
∆
.
/
ℎ
(
4
)
DA
=
∆
H
d
o
w
n
.
C
h
-
Ma
in
ten
an
ce
s
av
ed
DA
≈
(
−
)
+
∆
.
/
ℎ
(
5
)
-
T
h
r
o
u
g
h
p
u
t v
alu
e
∆
A
=
∆
H
d
o
w
n
/
T
p
l
an
n
ed
,
∆
OEE
≈
∆
A
-
No
m
in
al
p
r
o
d
u
ctio
n
g
ain
DA
=
N
omin
a
l
r
a
te
.
∆
OEE
.
H
r
un
.
M
a
r
gin
(
DA
/
unit
)
(
6
)
Year
-
1
R
OI
(
DA)
:
R
OI
=
DA
D
o
wnt
im
e
+
DA
Main
t
en
ac
e
+
DA
T
h
r
o
u
g
h
pu
t
−
Pr
o
ject
Co
s
t
(
DA
)
Pr
o
jec
t
Co
s
t
(
DA
)
(
7
)
w
h
er
e
A
d
e
n
o
tes
av
ailab
ilit
y
;
T
_
p
la
n
n
ed
d
en
o
tes
p
lan
n
ed
o
p
e
r
atin
g
tim
e;
T
_
d
o
wn
t
im
e
d
en
o
tes
t
o
tal
d
o
wn
tim
e
(
p
lan
n
e
d
a
n
d
u
n
p
lan
n
ed
)
;
MT
B
F
d
en
o
tes
m
ea
n
tim
e
b
etwe
en
f
ailu
r
es;
MT
T
R
d
en
o
tes
m
ea
n
tim
e
to
r
ep
air
;
OE
E
d
en
o
tes
o
v
er
al
l
eq
u
ip
m
en
t
ef
f
ec
tiv
e
n
ess
;
n
_
(
f
ailu
r
es
av
o
id
ed
)
d
en
o
tes
th
e
n
u
m
b
er
o
f
f
ailu
r
es
p
r
ev
en
ted
b
y
t
h
e
ML
m
o
d
e
l;
C
_
C
M
d
en
o
tes
co
r
r
ec
tiv
e
m
ain
ten
a
n
ce
c
o
s
t
p
er
f
ailu
r
e;
C
_
PM
d
en
o
tes
p
r
ev
en
tiv
e
m
ain
ten
a
n
ce
co
s
t
p
er
in
ter
v
e
n
tio
n
;
Δ
MT
T
R
d
en
o
tes
r
ed
u
ctio
n
in
r
ep
air
tim
e;
C
_
(
lab
o
r
/
h
)
d
en
o
tes
lab
o
r
co
s
t
p
er
h
o
u
r
;
Δ
H_
d
o
wn
d
en
o
tes
d
o
wn
tim
e
h
o
u
r
s
a
v
o
i
d
ed
;
C
_
h
d
en
o
tes
co
s
t
p
e
r
h
o
u
r
o
f
d
o
wn
tim
e;
Δ
A
d
en
o
tes
im
p
r
o
v
em
en
t
in
av
ai
lab
ilit
y
;
Δ
OE
E
d
en
o
tes
im
p
r
o
v
em
en
t
in
OE
E
;
H_
r
u
n
d
e
n
o
tes
to
tal
o
p
er
atin
g
h
o
u
r
s
a
f
ter
im
p
r
o
v
em
en
t;
No
m
in
al
r
ate
d
en
o
tes
n
o
m
in
al
p
r
o
d
u
ctio
n
r
ate;
Ma
r
g
in
(
DA/u
n
it)
d
en
o
tes
p
r
o
f
it
m
ar
g
in
p
er
u
n
it
p
r
o
d
u
ce
d
;
DA
d
en
o
tes
ec
o
n
o
m
ic
g
ain
e
x
p
r
e
s
s
ed
in
Alg
er
ian
Din
ar
s
;
an
d
R
OI
d
en
o
tes
r
etu
r
n
o
n
in
v
estme
n
t.
4.
ST
AND
ARDS
CO
M
P
L
I
AN
CE
S
T
RA
T
E
G
Y
Go
al:
E
n
s
u
r
e
an
i
n
d
u
s
tr
ial
M
L
s
o
lu
tio
n
th
at
is
in
ter
o
p
er
ab
le,
tr
ac
ea
b
le,
an
d
au
d
itab
le
b
y
alig
n
in
g
with
co
r
e
s
tan
d
ar
d
s
[
1
4
]
,
[
1
5
]
.
a.
R
ef
er
en
ce
f
r
am
ewo
r
k
s
to
f
o
llo
w
−
I
n
ter
n
atio
n
al
Or
g
a
n
izatio
n
f
o
r
Stan
d
ar
d
izatio
n
–
Stan
d
ar
d
1
3
3
7
4
(
I
SO
-
1
3
3
7
4
)
(
c
o
n
d
itio
n
m
o
n
ito
r
in
g
)
:
C
h
ain
f
u
n
ctio
n
s
a
cq
u
is
itio
n
→
d
ata
q
u
ality
/p
r
ep
r
o
ce
s
s
in
g
→
s
tate
d
etec
tio
n
→
h
ea
lth
a
s
s
es
s
m
en
t
→
p
r
o
g
n
o
s
tics
→
ad
v
is
o
r
y
/alar
m
s
.
−
I
n
ter
n
atio
n
al
E
lectr
o
tech
n
ical
C
o
m
m
is
s
io
n
Stan
d
ar
d
(
IEC
)
6
2
2
6
4
/
I
n
ter
n
atio
n
al
So
ciety
o
f
Au
to
m
atio
n
Stan
d
ar
d
9
5
(
I
SA
-
95
)
(
en
t
er
p
r
is
e
–
co
n
tr
o
l
in
te
g
r
atio
n
)
:
Place
co
m
p
o
n
e
n
ts
ac
r
o
s
s
l
ev
els
0
–
4
(
s
en
s
o
r
s
/
p
r
o
g
r
am
m
ab
le
lo
g
ic
co
n
tr
o
ller
(
PLC
)
/
s
u
p
er
v
is
o
r
y
co
n
tr
o
l
an
d
d
ata
ac
q
u
is
itio
n
(
SC
ADA
)
→
m
an
u
f
ac
tu
r
in
g
e
x
ec
u
tio
n
s
y
s
tem
(
MES
)
/
e
dge
→
en
ter
p
r
is
e
r
eso
u
r
ce
p
lan
n
in
g
(
E
R
P
)
/
b
u
s
in
es
s
in
tellig
en
ce
(
B
I
)
with
o
p
e
n
p
latf
o
r
m
c
o
m
m
u
n
icatio
n
s
–
u
n
if
ied
a
r
ch
itectu
r
e
(
OPC
UA
)
/
m
ess
ag
e
q
u
eu
in
g
telem
etr
y
tr
a
n
s
p
o
r
t (
MQ
T
T
)
in
ter
f
ac
es.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Po
w
E
lec
&
Dr
i Sy
s
t
I
SS
N:
2088
-
8
6
9
4
A
p
p
lica
tio
n
o
f m
a
ch
in
e
le
a
r
n
in
g
fo
r
p
r
o
d
u
ctio
n
o
p
timiz
a
tio
n
a
n
d
…
(
La
kh
d
a
r
i La
h
ce
n
)
769
b.
Min
im
u
m
im
p
lem
en
tatio
n
r
eq
u
ir
em
en
ts
−
Data
(
g
o
v
er
n
a
n
ce
)
:
C
atalo
g
an
d
lin
ea
g
e
(
s
en
s
o
r
→
f
ea
tu
r
e
→
p
r
ed
ictio
n
)
,
v
e
r
s
io
n
ed
s
ch
e
m
as,
q
u
ality
co
n
tr
o
ls
(
co
m
p
leten
ess
,
f
r
esh
n
ess
,
d
r
if
t)
,
r
o
le
-
b
ased
ac
ce
s
s
co
n
tr
o
l
(
R
B
AC
)
,
tr
an
s
p
o
r
t
lay
er
s
ec
u
r
ity
(
TLS
)
,
en
cr
y
p
tio
n
at
r
est,
r
ete
n
tio
n
p
o
licies,
an
d
au
d
it lo
g
s
.
−
Mo
d
els
(
ML
g
o
v
e
r
n
an
ce
)
:
Mo
d
el
r
e
g
is
tr
y
(
I
D/v
e
r
s
io
n
/
d
ata/lim
its
)
,
r
ep
r
o
d
u
cib
ilit
y
(
ar
tifa
cts
&
co
n
f
ig
s
)
,
ex
p
lain
ab
ilit
y
(
f
ea
tu
r
e
im
p
o
r
tan
ce
/SHAP
f
o
r
r
an
d
o
m
f
o
r
est
(
RF
)
/
g
r
ad
ien
t
b
o
o
s
tin
g
d
ec
is
io
n
tr
ee
(
GB
DT
)
,
d
o
cu
m
e
n
ted
th
r
esh
o
ld
s
with
h
y
s
ter
esis
.
−
Valid
atio
n
(
I
Q/OQ/PQ)
:
I
Q:
en
v
ir
o
n
m
en
ts
an
d
OPC
UA/
MQ
T
T
co
n
n
ec
tiv
ity
;
OQ:
o
f
f
lin
e
test
s
with
ac
ce
p
tan
ce
cr
iter
ia
(
F1
/AUC
o
r
R
MSE
/MAE
/R
²)
;
an
d
PQ:
o
n
lin
e
p
ilo
t
(
r
ea
d
-
o
n
ly
)
,
th
e
n
ca
n
ar
y
5
–
1
0
% with
clea
r
r
o
llb
ac
k
cr
iter
i
a.
−
Op
er
atio
n
s
(
lig
h
tweig
h
t
m
ac
h
in
e
lear
n
in
g
o
p
er
atio
n
s
(
ML
O
p
s
)
):
C
o
n
tin
u
o
u
s
m
o
n
ito
r
in
g
(
d
ata/m
etr
ic
d
r
if
t,
laten
cy
,
aler
t r
ates)
,
s
ch
e
d
u
led
r
ec
alib
r
atio
n
/r
etr
ain
in
g
,
f
o
r
m
al
ch
an
g
e
co
n
tr
o
l (
r
ev
iew
,
r
eg
r
ess
io
n
test
s
,
ap
p
r
o
v
al,
v
e
r
s
io
n
ed
r
ele
ase)
.
c.
R
eq
u
ir
ed
d
eliv
er
a
b
les
−
Data
m
an
ag
em
en
t
p
lan
,
I
Q/
OQ/PQ
d
o
s
s
ier
(
p
r
o
to
co
ls
&
r
esu
lts
)
,
m
o
d
el
r
eg
is
tr
y
,
SOPs
(
alar
m
p
r
o
ce
d
u
r
e,
d
eg
r
ad
e
d
/r
o
llb
ac
k
m
o
d
e)
,
d
ash
b
o
a
r
d
s
,
an
d
au
d
it l
o
g
s
.
5.
M
E
T
H
O
DO
L
O
G
Y
a.
Simu
latio
n
-
b
ased
d
ata
g
en
er
at
io
n
I
n
th
e
ab
s
en
ce
o
f
h
is
to
r
ical
p
l
an
t
d
ata,
we
g
en
e
r
ated
th
r
ee
s
y
n
th
etic
d
atasets
to
em
u
late
r
ea
lis
ti
c
s
ce
n
ar
io
s
.
T
h
r
ee
d
atasets
wer
e
cr
ea
ted
:
−
Pro
d
u
ctio
n
d
ataset:
I
n
clu
d
es
f
u
r
n
ac
e
tem
p
er
at
u
r
e,
p
r
ess
u
r
e,
p
r
o
d
u
ctio
n
r
ate,
en
e
r
g
y
c
o
n
s
u
m
p
tio
n
,
an
d
m
ea
s
u
r
ed
ir
o
n
co
n
ten
t
(
%).
T
h
ese
p
ar
am
eter
s
ar
e
cr
itical
f
o
r
m
etallu
r
g
ical
p
r
o
ce
s
s
es,
wh
er
e
s
m
all
d
ev
iatio
n
s
ca
n
s
ig
n
if
ican
tl
y
i
m
p
ac
t f
in
al
q
u
ality
[
1
6
]
‒
[
1
8
]
.
−
Ma
in
ten
an
ce
d
ataset:
I
n
clu
d
es
v
ib
r
atio
n
lev
el
(
m
m
/s
)
,
m
ac
h
in
e
tem
p
er
atu
r
e
(
°C
)
,
an
d
a
b
in
ar
y
f
ailu
r
e
s
tate
(
0
=
n
o
r
m
al,
1
=
f
ailu
r
e)
.
T
h
ese
v
ar
iab
les
ar
e
ty
p
ical
in
d
icato
r
s
o
f
m
ec
h
an
ical
a
n
d
th
er
m
al
s
tr
ess
in
in
d
u
s
tr
ial
eq
u
i
p
m
en
t
[
1
7
]
.
−
T
r
an
s
p
o
r
t
d
ataset:
I
n
clu
d
es
s
h
ip
m
en
t
to
n
n
ag
e,
tr
a
n
s
p
o
r
t
d
is
tan
ce
,
co
s
t
(
in
DZ
D)
,
an
d
tr
av
el
tim
e
(
h
o
u
r
s
)
.
T
r
an
s
p
o
r
t lo
g
is
tics
r
ep
r
esen
t a
m
ajo
r
c
o
s
t f
ac
to
r
in
lar
g
e
-
s
ca
le
in
d
u
s
tr
ial
o
p
e
r
atio
n
s
[
1
6
]
‒
[
1
8
]
.
b.
Ma
ch
in
e
lear
n
in
g
m
o
d
els
Fiv
e
wid
ely
u
s
ed
ML
alg
o
r
ith
m
s
wer
e
s
elec
ted
f
o
r
ev
alu
atio
n
:
−
R
an
d
o
m
f
o
r
est:
An
en
s
em
b
le
m
eth
o
d
u
s
in
g
m
u
ltip
le
d
ec
is
io
n
tr
ee
s
to
im
p
r
o
v
e
p
r
e
d
ictiv
e
a
cc
u
r
ac
y
an
d
co
n
tr
o
l o
v
er
f
itti
n
g
.
−
XGBo
o
s
t: An
ef
f
icien
t g
r
ad
ie
n
t b
o
o
s
tin
g
al
g
o
r
ith
m
o
p
tim
iz
ed
f
o
r
p
er
f
o
r
m
an
ce
an
d
s
p
ee
d
.
−
L
ig
h
tGB
M:
A
g
r
ad
ien
t
b
o
o
s
t
in
g
f
r
a
m
ewo
r
k
b
ased
o
n
d
ec
is
io
n
tr
ee
alg
o
r
ith
m
s
,
d
esig
n
ed
f
o
r
h
ig
h
ef
f
icien
cy
an
d
lo
w
m
em
o
r
y
u
s
ag
e.
−
C
atB
o
o
s
t:
A
g
r
ad
ien
t
b
o
o
s
tin
g
alg
o
r
ith
m
p
ar
ticu
la
r
ly
ef
f
ec
tiv
e
with
ca
teg
o
r
ical
f
ea
tu
r
es
a
n
d
r
e
d
u
cin
g
o
v
er
f
itti
n
g
.
−
Mu
ltil
ay
er
p
er
ce
p
tr
o
n
(
ML
P):
A
f
ee
d
f
o
r
war
d
ar
tific
ial
n
eu
r
al
n
etwo
r
k
ca
p
ab
le
o
f
m
o
d
elin
g
co
m
p
lex
non
-
lin
ea
r
r
elatio
n
s
h
ip
s
[
1
4
]
.
−
Hy
b
r
id
e
R
F+NN
(
s
tack
in
g
)
:
T
h
e
“RF
+N
N”
ap
p
r
o
ac
h
c
o
m
b
i
n
es
a
r
an
d
o
m
f
o
r
est
(
R
F)
as
a
b
ase
lear
n
er
an
d
a
m
u
ltil
ay
er
p
er
ce
p
tr
o
n
(
ML
P)
as a
m
eta
-
m
o
d
el
T
h
e
g
l
o
b
a
l
w
o
r
k
f
l
o
w
o
f
t
h
e
p
r
o
p
o
s
e
d
M
a
c
h
i
n
e
l
e
a
r
n
i
n
g
-
b
a
s
ed
o
p
t
i
m
i
z
a
t
i
o
n
a
p
p
r
o
a
c
h
i
s
d
e
p
i
c
t
e
d
i
n
F
i
g
u
r
e
3
.
I
t
o
u
tlin
es
th
e
s
eq
u
en
tial
s
tep
s
f
r
o
m
s
y
n
th
etic
d
ataset
g
en
er
atio
n
to
m
o
d
el
ev
al
u
atio
n
an
d
in
ter
p
r
etatio
n
,
alig
n
in
g
with
b
o
th
p
r
ed
ictiv
e
m
ain
ten
an
ce
an
d
p
r
o
d
u
cti
o
n
q
u
ality
im
p
r
o
v
em
en
t o
b
jectiv
es [
1
9
]
.
c.
E
v
alu
atio
n
m
etr
ics
Fo
r
class
if
icatio
n
(
p
r
ed
ictiv
e
m
ain
ten
an
ce
)
[
2
0
]
‒
[
2
2
]
:
−
Acc
u
r
ac
y
:
p
r
o
p
o
r
tio
n
o
f
co
r
r
e
ct
p
r
ed
ictio
n
s
=
+
+
+
+
(
8
)
−
Pre
cisi
o
n
:
p
r
o
p
o
r
tio
n
o
f
co
r
r
e
ctly
p
r
ed
icted
p
o
s
itiv
es a
m
o
n
g
all
p
r
ed
icted
p
o
s
itiv
es
Pr
e
c
isio
n
=
TP
TP
+
FP
(
9
)
−
R
ec
all
:
p
r
o
p
o
r
tio
n
o
f
c
o
r
r
ec
tly
p
r
ed
icted
p
o
s
itiv
es a
m
o
n
g
all
ac
tu
al
p
o
s
itiv
es
R
e
c
a
l
l
=
TP
TP
+
FN
(
1
0
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
6
9
4
I
n
t J Po
w
E
lec
&
Dr
i Sy
s
t
,
Vo
l.
1
7
,
No
.
1
,
Ma
r
c
h
20
2
6
:
765
-
776
770
−
F1
-
s
co
r
e
:
h
ar
m
o
n
ic
m
ea
n
o
f
p
r
ec
is
io
n
an
d
r
ec
all
F1
−
s
c
or
e
=
2
×
Pr
ecis
i
o
n
X
Recal
l
Pr
ecis
i
o
n
+
Recal
l
(
1
1
)
−
AUC
-
R
O
C
:
ar
ea
u
n
d
er
th
e
r
ec
eiv
er
o
p
e
r
atin
g
ch
ar
a
cter
is
tic
cu
r
v
e,
in
d
icatin
g
th
e
m
o
d
el’
s
d
is
cr
im
in
atio
n
ab
ilit
y
.
AUC
=
∫
TPR
(
F
PR
)
d
(
F
PR
)
1
0
(
1
2
)
w
ith
TPR
=
TP
TP
+
FN
,
F
PR
=
FP
FP
+
TN
w
h
er
e
:
T
P (
tr
u
e
p
o
s
itiv
es
)
,
T
N
(
tr
u
e
n
e
g
ativ
es
)
,
FP
(
f
alse p
o
s
itiv
es
)
,
an
d
FN (
f
alse n
eg
ativ
e
s
)
.
Fo
r
r
eg
r
ess
io
n
(
q
u
ality
p
r
ed
ict
io
n
)
[
2
3
]‒
[
2
5
]
:
−
R
o
o
t
m
ea
n
s
q
u
ar
e
d
er
r
o
r
(
R
M
SE)
:
m
ea
s
u
r
es p
r
ed
ictio
n
e
r
r
o
r
m
ag
n
itu
d
e
R
M
SE
=
√
1
n
∑
(
yi
−
y
̂
i
)
2
n
i
=
1
(
1
3
)
−
Me
an
ab
s
o
lu
te
er
r
o
r
(
MA
E
)
:
a
v
er
ag
e
o
f
ab
s
o
lu
te
p
r
ed
ictio
n
er
r
o
r
s
MAE
=
1
n
∑
|
y
i
−
y
̂
i
|
n
i
=
1
(
1
4
)
−
C
o
ef
f
icien
t o
f
d
eter
m
in
atio
n
(
R
²)
:
m
ea
s
u
r
es th
e
p
r
o
p
o
r
tio
n
o
f
v
ar
ian
ce
e
x
p
lain
ed
b
y
t
h
e
m
o
d
el.
R
2
=
1
−
∑
(
−
̂
)
2
=
1
∑
(
−
̅
)
2
=
1
(
1
5
)
W
ith
y
̂
i
is
t
h
e
p
r
ed
icted
v
alu
es
;
yi
is
r
ea
l
v
alu
es
,
an
d
̅
: is th
e
av
e
r
ag
e
o
f
th
e
ac
tu
al
v
al
u
es
.
Fig
u
r
e
3
.
Me
th
o
d
o
lo
g
ical
f
lo
w
ch
ar
t o
f
t
h
e
ML
-
b
ased
i
n
d
u
s
tr
i
al
o
p
tim
izatio
n
f
r
a
m
ewo
r
k
d.
Mo
d
el
tr
ain
in
g
an
d
v
alid
atio
n
Data
wer
e
s
p
lit
in
to
tr
ain
in
g
(
7
0
%)
an
d
test
(
3
0
%)
s
ets.
K
-
f
o
ld
cr
o
s
s
-
v
alid
atio
n
was
ap
p
lied
to
en
s
u
r
e
r
o
b
u
s
tn
ess
.
SHap
ley
Ad
d
itiv
e
ex
Plan
atio
n
s
(
SHAP
)
an
aly
s
is
was
u
s
ed
to
in
ter
p
r
et
th
e
im
p
ac
t
o
f
ea
ch
f
ea
tu
r
e
o
n
m
o
d
el
p
r
ed
ictio
n
s
[
2
6
]
‒
[
2
8
]
.
Fig
u
r
e
4
illu
s
tr
ates
th
e
g
lo
b
al
s
im
u
latio
n
an
d
Ma
ch
in
e
lear
n
in
g
wo
r
k
f
lo
w
ad
o
p
ted
in
th
is
s
tu
d
y
.
T
h
e
p
r
o
ce
s
s
b
eg
in
s
with
th
e
g
en
er
atio
n
an
d
p
r
ep
r
o
ce
s
s
in
g
o
f
s
y
n
th
etic
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Po
w
E
lec
&
Dr
i Sy
s
t
I
SS
N:
2088
-
8
6
9
4
A
p
p
lica
tio
n
o
f m
a
ch
in
e
le
a
r
n
in
g
fo
r
p
r
o
d
u
ctio
n
o
p
timiz
a
tio
n
a
n
d
…
(
La
kh
d
a
r
i La
h
ce
n
)
771
d
atasets
co
v
er
in
g
p
r
o
d
u
ctio
n
,
m
ain
ten
an
ce
,
an
d
tr
an
s
p
o
r
t o
p
er
atio
n
s
.
T
h
ese
d
atasets
ar
e
th
en
u
s
ed
to
tr
ain
a
n
d
ev
alu
ate
f
iv
e
Ma
ch
in
e
lear
n
in
g
alg
o
r
ith
m
s
f
o
r
class
if
icatio
n
a
n
d
r
eg
r
ess
io
n
task
s
.
T
h
e
p
ip
elin
e
in
teg
r
ates
p
er
f
o
r
m
an
ce
a
n
aly
s
is
,
in
ter
p
r
etab
ilit
y
tech
n
iq
u
es
(
f
ea
tu
r
e
i
m
p
o
r
tan
ce
an
d
SHAP),
an
d
v
is
u
al
an
aly
tics
to
ex
tr
ac
t
ac
tio
n
ab
le
in
s
ig
h
ts
.
Fi
n
ally
,
all
r
esu
lts
ar
e
co
m
p
iled
in
a
s
tr
u
ctu
r
ed
f
o
r
m
at
f
o
r
in
c
lu
s
io
n
in
in
d
u
s
tr
ial
d
ec
is
io
n
-
m
ak
in
g
an
d
s
cien
tific
r
ep
o
r
tin
g
.
Data
s
p
litt
in
g
:
−
C
h
r
o
n
o
lo
g
ical
s
p
lit to
a
v
o
id
le
ak
ag
e:
tr
ain
/v
alid
atio
n
/tes
t
=
7
0
/1
5
/1
5
(
o
r
7
0
/3
0
with
C
V
o
n
th
e
tr
ain
p
ar
t)
.
−
Valid
atio
n
:
T
im
eSer
iesS
p
lit
(
k
=5
)
f
o
r
tim
e
-
o
r
d
e
r
ed
s
er
ies
(
p
r
o
d
u
ctio
n
/tra
n
s
p
o
r
t)
a
n
d
St
r
atif
ied
K
-
Fo
ld
(
k
=5
)
f
o
r
b
in
ar
y
m
ain
ten
an
ce
lab
els.
−
T
h
e
test
s
et
r
em
ain
s
u
n
to
u
ch
e
d
u
n
til f
in
al
r
ep
o
r
tin
g
.
Fig
u
r
e
4
.
Flo
wch
ar
t
o
f
th
e
s
im
u
latio
n
an
d
m
ac
h
in
e
lea
r
n
in
g
o
p
tim
izatio
n
p
ip
elin
e
,
f
r
o
m
d
ataset
in
p
u
t to
r
esu
lt e
x
p
o
r
t
6.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
6
.
1
.
Ris
k
o
f
bia
s
f
ro
m
s
y
nthet
ic
da
t
a
a
nd
m
it
i
g
a
t
io
n
Natu
r
e
o
f
th
e
b
ias.
Sy
n
th
etic
d
atasets
m
ay
u
n
d
er
-
r
e
p
r
esen
t
ce
r
tain
r
eg
im
es
(
tr
an
s
ien
ts
,
r
ar
e
f
ailu
r
es),
s
im
p
lify
co
r
r
elatio
n
s
(
n
o
n
-
lin
e
ar
ities
,
cr
o
s
s
-
in
ter
ac
tio
n
s
)
,
an
d
o
v
er
esti
m
ate
s
tatio
n
ar
ity
(
n
o
is
e/d
r
if
t
lo
wer
th
an
in
th
e
f
ield
)
.
T
h
is
ca
n
y
ield
o
p
tim
is
tic
p
er
f
o
r
m
an
ce
(
h
ig
h
AUC/
F1
)
an
d
o
v
er
f
itti
n
g
to
s
im
u
latio
n
ass
u
m
p
tio
n
s
.
Mitig
atio
n
m
ea
s
u
r
es (
ap
p
lied
/
r
ec
o
m
m
en
d
ed
)
:
-
Do
m
ain
r
an
d
o
m
izatio
n
:
v
a
r
y
n
o
is
e
am
p
litu
d
es,
s
en
s
o
r
g
ain
/o
f
f
s
et,
o
p
e
r
atin
g
r
e
g
im
es
(
th
r
o
u
g
h
p
u
t,
tem
p
er
atu
r
es),
a
n
d
in
ject
o
u
tlier
s
/m
is
s
in
g
v
alu
es →
r
ed
u
ce
s
o
v
er
f
itti
n
g
t
o
id
ea
lized
d
is
tr
ib
u
tio
n
s
.
-
Strict
tem
p
o
r
al
v
alid
atio
n
:
c
h
r
o
n
o
lo
g
ical
s
p
lits
+
T
im
eSer
i
esS
p
lit
(
p
r
ev
en
ts
leak
ag
e
)
a
n
d
r
ep
o
r
t
m
ea
n
±
s
td
ac
r
o
s
s
f
o
ld
s
.
-
Sen
s
itiv
ity
&
s
tr
ess
test
s
:
r
e
-
ev
alu
ate
m
etr
ics
with
±
1
0
–
2
0
%
ch
an
g
es
i
n
s
am
p
lin
g
r
ate,
win
d
o
w
s
ize,
a
n
d
n
o
is
e
in
ten
s
ity
.
-
Ab
latio
n
s
&
p
ar
s
im
o
n
y
:
s
elec
tiv
ely
r
em
o
v
e
v
ar
iab
les
(
v
ib
r
atio
n
a
n
d
tem
p
er
atu
r
e)
t
o
v
er
if
y
m
o
d
el
d
ep
en
d
e
n
ce
; u
s
e
p
en
alties/
ea
r
l
y
s
to
p
p
in
g
f
o
r
b
o
o
s
ter
s
/ML
P.
-
C
alib
r
atio
n
&
th
r
esh
o
ld
s
:
ca
l
ib
r
atio
n
cu
r
v
es
an
d
v
alid
atio
n
-
b
ased
th
r
esh
o
ld
in
g
(
Yo
u
d
e
n
/F1
)
,
th
en
l
o
ck
th
r
esh
o
ld
s
b
ef
o
r
e
test
in
g
.
-
Path
to
r
ea
l
d
ata:
o
f
f
lin
e
r
ep
la
y
o
f
SC
ADA
tag
s
,
th
en
r
ea
d
-
o
n
ly
POC
(
laten
cy
/alar
m
s
)
an
d
ca
n
ar
y
r
o
llo
u
t
(5
–
1
0
%)
with
r
o
llb
ac
k
cr
iter
ia
.
6
.
2
.
Cla
s
s
if
ica
t
io
n r
esu
lt
s
K
e
y
i
n
s
i
g
h
ts
:
i
)
V
i
b
r
a
ti
o
n
a
n
d
m
a
c
h
i
n
e
t
e
m
p
e
r
a
t
u
r
e
w
e
r
e
t
h
e
t
o
p
p
r
e
d
i
c
t
o
r
s
o
f
f
a
i
l
u
r
e
;
a
n
d
i
i
)
S
H
AP
a
n
a
l
y
s
is
c
o
n
f
i
r
m
e
d
t
h
e
i
r
s
t
r
o
n
g
i
n
f
l
u
e
n
c
e
o
n
c
l
a
s
s
i
f
i
ca
t
i
o
n
o
u
t
p
u
t
.
F
i
g
u
r
e
s
5
t
o
1
5
o
f
f
e
r
a
n
i
n
t
e
g
r
a
t
e
d
v
i
s
u
a
l
a
s
s
e
s
s
m
e
n
t
o
f
t
h
e
c
l
as
s
i
f
i
c
at
i
o
n
r
e
s
u
l
ts
o
b
t
a
i
n
e
d
f
r
o
m
al
l
te
s
te
d
m
o
d
e
l
s
.
F
i
g
u
r
e
s
5
t
o
9
d
i
s
p
l
a
y
t
h
e
c
o
m
p
a
r
a
ti
v
e
a
c
c
u
r
a
c
y
,
p
r
e
c
i
s
i
o
n
,
r
e
c
a
ll
,
F
1
-
s
c
o
r
e
,
a
n
d
A
U
C
m
e
t
r
ic
s
,
c
l
ea
r
l
y
s
h
o
w
i
n
g
t
h
e
d
o
m
i
n
a
n
c
e
o
f
t
r
e
e
-
b
a
s
e
d
m
e
t
h
o
d
s
(
r
a
n
d
o
m
f
o
r
e
s
t
,
X
GB
o
o
s
t
,
L
i
g
h
t
G
B
M
,
C
at
B
o
o
s
t
,
a
n
d
R
F
+N
N
)
o
v
e
r
t
h
e
N
e
u
r
a
l
N
et
w
o
r
k
b
a
s
e
l
i
n
e
.
F
i
g
u
r
e
1
0
p
r
esen
ts
th
e
R
OC
cu
r
v
es,
wh
er
e
en
s
em
b
le
m
o
d
els
ac
h
ie
v
e
n
ea
r
-
p
er
f
ec
t
class
s
ep
ar
atio
n
.
T
h
e
co
n
f
u
s
io
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
6
9
4
I
n
t J Po
w
E
lec
&
Dr
i Sy
s
t
,
Vo
l.
1
7
,
No
.
1
,
Ma
r
c
h
20
2
6
:
765
-
776
772
m
atr
ix
in
Fig
u
r
e
1
1
f
u
r
th
e
r
v
alid
ates
th
e
r
an
d
o
m
f
o
r
est
m
o
d
el’
s
p
er
f
o
r
m
an
ce
,
r
e
v
ea
lin
g
ze
r
o
f
alse
p
o
s
itiv
es
an
d
ze
r
o
f
alse
n
eg
ativ
es
in
t
h
e
s
im
u
lated
d
ataset.
Fin
ally
,
Fig
u
r
e
1
2
illu
s
tr
ates
th
e
r
an
d
o
m
f
o
r
est
f
ea
tu
r
e
-
im
p
o
r
tan
ce
p
r
o
f
ile,
i
d
en
tify
in
g
v
ib
r
atio
n
lev
el
an
d
m
ac
h
i
n
e
tem
p
er
atu
r
e
as th
e
m
o
s
t im
p
ac
tf
u
l p
r
ed
icto
r
s
.
Fig
u
r
e
5
.
C
o
m
p
a
r
ativ
e
ac
cu
r
a
cy
Fig
u
r
e
6
.
C
o
m
p
a
r
ativ
e
p
r
ec
is
io
n
Fig
u
r
e
7
.
C
o
m
p
a
r
ativ
e
r
ec
all
Fig
u
r
e
8
.
C
o
m
p
a
r
ativ
e
F1
-
Sco
r
e
Fig
u
r
e
9
.
C
o
m
p
a
r
ativ
e
AUC
Fig
u
r
e
1
0
.
R
OC
cu
r
v
es f
o
r
cla
s
s
if
icatio
n
m
o
d
els
6
.
3
.
Reg
re
s
s
io
n r
esu
lt
s
Fig
u
r
es
1
3
to
1
5
p
r
esen
t
th
e
c
o
m
p
ar
ativ
e
r
e
g
r
ess
io
n
p
er
f
o
r
m
an
ce
o
f
all
ev
al
u
ated
m
o
d
el
s
b
ased
o
n
R
MSE
,
MA
E
,
an
d
R
²
m
etr
ics.
T
h
e
r
esu
lts
s
h
o
w
th
at
tr
ee
-
b
ased
en
s
em
b
le
m
eth
o
d
s
(
r
an
d
o
m
f
o
r
est,
XGBo
o
s
t,
L
ig
h
tGB
M,
C
atB
o
o
s
t)
co
n
s
is
t
en
tly
o
u
tp
er
f
o
r
m
th
e
Neu
r
al
Netwo
r
k
b
aselin
e.
L
ig
h
tGB
M
ac
h
iev
es
th
e
lo
west
R
MSE
an
d
lo
west
MA
E
,
w
h
ile
also
o
b
tain
in
g
th
e
h
ig
h
e
s
t
R
²,
in
d
icatin
g
b
etter
p
r
ed
i
ctio
n
s
tab
ilit
y
an
d
r
ed
u
ce
d
e
r
r
o
r
v
ar
iab
ilit
y
.
C
o
n
v
er
s
ely
,
th
e
s
tan
d
alo
n
e
Neu
r
al
Netwo
r
k
s
h
o
ws
th
e
wea
k
est
p
er
f
o
r
m
an
ce
ac
r
o
s
s
all
m
etr
ics,
co
n
f
ir
m
in
g
its
lim
ited
s
u
itab
ilit
y
f
o
r
th
is
ty
p
e
o
f
in
d
u
s
tr
ial
r
eg
r
ess
io
n
task
.
T
h
e
h
y
b
r
id
R
F+NN
m
o
d
el
ac
h
iev
es
ac
ce
p
ta
b
le
r
es
u
lts
b
u
t
d
o
es
n
o
t
s
u
r
p
ass
th
e
b
est
in
d
iv
id
u
al
tr
ee
-
b
ased
m
o
d
els,
s
u
g
g
esti
n
g
t
h
at
h
y
b
r
id
izatio
n
o
f
f
er
s
o
n
ly
m
ar
g
in
al
b
en
ef
it f
o
r
th
is
r
eg
r
ess
io
n
p
r
o
b
lem
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Po
w
E
lec
&
Dr
i Sy
s
t
I
SS
N:
2088
-
8
6
9
4
A
p
p
lica
tio
n
o
f m
a
ch
in
e
le
a
r
n
in
g
fo
r
p
r
o
d
u
ctio
n
o
p
timiz
a
tio
n
a
n
d
…
(
La
kh
d
a
r
i La
h
ce
n
)
773
Fig
u
r
e
1
1
.
C
o
n
f
u
s
io
n
m
atr
ix
f
o
r
r
an
d
o
m
f
o
r
est
Fig
u
r
e
1
2
.
Featu
r
e
im
p
o
r
tan
ce
r
an
d
o
m
f
o
r
est
Fig
u
r
e
1
3
.
C
o
m
p
ar
ativ
e
R
MSE
Fig
u
r
e
1
4
.
C
o
m
p
ar
ativ
e
MA
E
Fig
u
r
e
1
5
.
C
o
m
p
ar
ativ
e
R
2
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
6
9
4
I
n
t J Po
w
E
lec
&
Dr
i Sy
s
t
,
Vo
l.
1
7
,
No
.
1
,
Ma
r
c
h
20
2
6
:
765
-
776
774
7.
CO
NCLU
SI
O
N
T
h
i
s
s
t
u
d
y
d
e
m
o
n
s
t
r
a
te
d
t
h
e
p
o
t
e
n
t
i
a
l
o
f
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
t
ec
h
n
i
q
u
e
s
f
o
r
b
o
t
h
p
r
e
d
i
c
t
i
v
e
m
a
i
n
t
e
n
a
n
c
e
a
n
d
p
r
o
d
u
c
t
i
o
n
q
u
a
l
i
t
y
o
p
t
i
m
iz
a
t
i
o
n
i
n
t
h
e
c
o
n
t
e
x
t
o
f
a
n
i
r
o
n
p
r
o
c
e
s
s
i
n
g
p
l
a
n
t
.
B
y
u
s
i
n
g
a
s
i
m
u
l
a
t
i
o
n
-
b
a
s
e
d
a
p
p
r
o
a
c
h
,
w
e
e
v
a
l
u
a
t
e
d
s
i
x
w
i
d
e
l
y
a
d
o
p
t
e
d
M
L
m
o
d
e
l
s
,
RF
,
X
G
B
o
o
s
t
,
L
i
g
h
t
G
B
M
,
C
a
t
B
o
o
s
t
,
f
e
e
d
f
o
r
w
a
r
d
n
e
u
r
a
l
n
e
t
w
o
r
k
,
a
n
d
an
R
F
+
N
N
o
n
3
s
y
n
t
h
e
t
i
c
d
a
t
a
s
e
t
s
r
e
p
r
e
s
e
n
t
i
n
g
p
r
o
d
u
c
t
i
o
n
,
m
a
i
n
t
e
n
a
n
c
e
,
a
n
d
t
r
a
n
s
p
o
r
t
p
r
o
c
e
s
s
e
s
.
T
h
e
r
esu
lts
o
f
th
e
class
if
icatio
n
task
s
,
aim
ed
at
p
r
ed
ict
in
g
eq
u
i
p
m
en
t
f
ailu
r
es,
r
ev
e
aled
th
at
en
s
em
b
le
tr
ee
-
b
ased
m
o
d
els
ac
h
iev
ed
n
ea
r
-
p
er
f
ec
t
p
r
e
d
ictiv
e
p
er
f
o
r
m
a
n
ce
,
with
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all
,
an
d
F1
-
s
co
r
e
v
alu
es
ex
ce
ed
i
n
g
0
.
9
9
,
a
n
d
AUC
v
alu
es
clo
s
e
to
1
.
Su
ch
p
er
f
o
r
m
an
ce
wa
s
ex
p
lain
ed
b
y
th
e
well
-
d
ef
in
ed
f
ailu
r
e
p
atter
n
s
i
n
th
e
s
im
u
lated
m
ain
te
n
an
ce
d
ata.
I
n
th
e
r
eg
r
ess
io
n
task
s
,
wh
ich
f
o
c
u
s
ed
o
n
p
r
ed
ictin
g
ir
o
n
co
n
ten
t
as
a
m
ea
s
u
r
e
o
f
p
r
o
d
u
ctio
n
q
u
ality
,
g
r
ad
ien
t
b
o
o
s
tin
g
alg
o
r
ith
m
s
s
u
ch
as
L
ig
h
tGB
M
an
d
C
atB
o
o
s
t
o
u
tp
er
f
o
r
m
e
d
o
th
er
m
o
d
els,
ac
h
iev
i
n
g
t
h
e
lo
west
R
MSE
an
d
MA
E
v
alu
es
an
d
th
e
h
ig
h
est
R
²
s
co
r
es (
ab
o
v
e
0
.
5
9
)
,
ev
e
n
in
th
e
p
r
esen
ce
o
f
s
im
u
lated
m
ea
s
u
r
em
en
t n
o
is
e.
T
h
e
r
an
d
o
m
f
o
r
est
alg
o
r
ith
m
a
ch
iev
ed
th
e
b
est
r
esu
lts
d
u
e
to
its
r
o
b
u
s
tn
ess
to
n
o
is
e,
ab
ilit
y
to
m
o
d
e
l
n
o
n
lin
ea
r
r
elatio
n
s
h
i
p
s
,
an
d
ef
f
ec
tiv
e
f
ea
tu
r
e
s
elec
tio
n
,
m
a
k
in
g
it
well
-
s
u
ited
f
o
r
in
d
u
s
tr
ial
tab
u
lar
d
ata.
T
h
e
h
y
b
r
id
R
F
+
ML
P
m
o
d
el
p
r
o
v
id
ed
o
n
ly
a
s
lig
h
t
im
p
r
o
v
em
en
t
s
in
ce
th
e
r
an
d
o
m
f
o
r
est
al
r
ea
d
y
r
ea
c
h
ed
n
ea
r
-
p
er
f
ec
t
p
er
f
o
r
m
a
n
ce
(
AUC
≈
1
.
0
)
.
T
h
is
m
ar
g
in
al
d
i
f
f
er
en
ce
ca
n
b
e
attr
ib
u
te
d
to
p
e
r
f
o
r
m
a
n
ce
s
atu
r
atio
n
an
d
in
f
o
r
m
atio
n
al
r
ed
u
n
d
an
c
y
,
as t
h
e
ML
P a
d
d
s
litt
le
v
alu
e
wh
en
th
e
d
ataset
is
alr
ea
d
y
well
s
ep
ar
ated
b
y
t
h
e
R
F.
B
ey
o
n
d
th
e
n
u
m
er
ical
p
er
f
o
r
m
an
ce
,
th
e
ex
p
lain
a
b
ilit
y
an
al
y
s
is
u
s
in
g
Featu
r
e
I
m
p
o
r
tan
ce
an
d
SHAP
p
r
o
v
id
e
d
v
alu
a
b
le
in
s
ig
h
ts
in
to
th
e
d
r
iv
in
g
f
ac
to
r
s
b
eh
i
n
d
th
e
p
r
ed
ictio
n
s
.
Fo
r
p
r
ed
ictiv
e
m
ain
ten
an
ce
,
v
ib
r
atio
n
lev
els
an
d
m
ac
h
in
e
t
em
p
er
atu
r
e
wer
e
id
e
n
tifie
d
as
th
e
d
o
m
in
an
t
p
r
ed
icto
r
s
o
f
f
ai
lu
r
e,
alig
n
in
g
with
r
ea
l
-
wo
r
ld
e
n
g
in
ee
r
in
g
k
n
o
wl
ed
g
e.
T
h
is
tr
an
s
p
ar
en
cy
is
ess
en
tial
f
o
r
in
d
u
s
tr
ial
ad
o
p
tio
n
,
as
it
s
u
p
p
o
r
ts
tr
u
s
t
in
th
e
m
o
d
els an
d
f
ac
ilit
ates d
ec
is
io
n
-
m
ak
in
g
b
y
e
n
g
in
ee
r
s
a
n
d
o
p
e
r
ato
r
s
.
W
h
ile
th
e
s
tu
d
y
is
b
ased
o
n
s
y
n
th
etic
d
ata,
its
m
eth
o
d
o
lo
g
y
is
d
esig
n
e
d
to
b
e
tr
a
n
s
f
er
a
b
le
to
r
ea
l
in
d
u
s
tr
ial
en
v
ir
o
n
m
en
ts
.
T
h
e
co
m
b
in
atio
n
o
f
r
o
b
u
s
t
ML
alg
o
r
ith
m
s
,
in
ter
p
r
etab
ilit
y
m
eth
o
d
s
,
an
d
a
well
-
s
tr
u
ctu
r
ed
ev
alu
atio
n
f
r
am
ew
o
r
k
ca
n
d
ir
ec
tly
s
u
p
p
o
r
t
r
ea
l
-
w
o
r
ld
d
ep
l
o
y
m
en
ts
o
n
ce
h
is
to
r
i
ca
l
o
p
er
atio
n
al
d
ata
b
ec
o
m
es
av
ailab
le.
T
h
e
ap
p
r
o
ac
h
r
em
ain
s
tr
a
n
s
f
er
ab
le
to
in
d
u
s
tr
ial
s
ettin
g
s
v
ia
SC
A
DA/I
I
o
T
in
teg
r
atio
n
(
OPC
UA/M
QT
T
)
an
d
ONN
X
p
ac
k
ag
i
n
g
at
th
e
ed
g
e,
wit
h
an
ex
p
licit
lin
k
b
etwe
en
tech
n
ical
m
etr
ics
a
n
d
p
lan
t
KPI
s
(
d
o
wn
tim
e,
m
ain
ten
an
ce
co
s
ts
,
th
r
o
u
g
h
p
u
t
in
DZ
D)
.
T
h
e
lo
g
ical
n
ex
t
s
tep
s
ar
e
to
r
ep
lay
SC
ADA
d
at
a
o
f
f
li
n
e,
r
u
n
a
r
ea
d
-
o
n
l
y
p
i
lo
t
to
ca
l
ib
r
a
te
t
h
r
es
h
o
l
d
s
a
n
d
l
ate
n
c
y
/
d
r
i
f
t,
e
x
ec
u
t
e
a
lim
ite
d
ca
n
a
r
y
d
ep
lo
y
m
e
n
t
with
r
o
llb
ac
k
cr
iter
ia,
an
d
t
h
e
n
in
d
u
s
tr
ialize
(
m
o
d
el
r
eg
is
tr
y
,
I
Q/OQ/PQ
v
alid
atio
n
,
o
p
er
at
o
r
SOPs
)
.
8.
L
I
M
I
T
AT
I
O
NS A
N
D
F
UT
U
RE
WO
RK
a.
L
im
itatio
n
s
o
f
s
y
n
th
etic
d
ata
Alth
o
u
g
h
th
is
s
tu
d
y
em
p
lo
y
s
s
im
u
latio
n
-
b
ased
d
atasets
,
it is
im
p
o
r
tan
t to
ac
k
n
o
wled
g
e
th
at
s
y
n
th
etic
d
ata
m
ay
n
o
t
f
u
lly
ca
p
tu
r
e
th
e
v
ar
iab
ilit
y
,
s
en
s
o
r
n
o
is
e,
an
d
r
ar
e
f
ailu
r
e
ev
en
ts
ty
p
ically
o
b
s
er
v
ed
in
r
ea
l
in
d
u
s
tr
ial
en
v
ir
o
n
m
en
ts
.
Su
ch
lim
itatio
n
s
ca
n
in
f
lu
en
ce
th
e
g
en
er
aliza
b
ilit
y
o
f
th
e
tr
ain
e
d
m
o
d
els.
T
h
er
ef
o
r
e,
f
u
tu
r
e
wo
r
k
will
f
o
cu
s
o
n
v
ali
d
atin
g
th
e
p
r
o
p
o
s
ed
m
eth
o
d
o
l
o
g
y
u
s
in
g
r
ea
l
o
p
e
r
atio
n
al
d
at
a
co
llected
f
r
o
m
th
e
B
ec
h
ar
ir
o
n
p
r
o
ce
s
s
in
g
f
ac
ilit
y
to
en
s
u
r
e
r
o
b
u
s
tn
ess
an
d
p
r
a
ctica
l r
eliab
ilit
y
.
b.
Hu
m
an
in
th
e
lo
o
p
an
d
o
p
er
at
o
r
tr
u
s
t
I
n
ad
d
itio
n
,
i
n
co
r
p
o
r
atin
g
h
u
m
an
ex
p
er
tis
e
in
to
th
e
d
ec
is
io
n
-
m
ak
in
g
p
r
o
ce
s
s
r
em
ain
s
cr
u
cial
f
o
r
th
e
s
u
cc
ess
f
u
l
in
d
u
s
tr
ial
d
ep
lo
y
m
en
t
o
f
m
ac
h
in
e
lear
n
in
g
s
y
s
tem
s
.
B
y
in
teg
r
atin
g
ex
p
lain
ab
il
ity
tech
n
iq
u
es
s
u
ch
as
SHA
P
an
d
f
ea
tu
r
e
im
p
o
r
ta
n
ce
an
aly
s
is
,
th
e
f
r
am
ewo
r
k
e
n
h
an
ce
s
tr
an
s
p
ar
en
c
y
an
d
s
u
p
p
o
r
ts
o
p
er
ato
r
tr
u
s
t.
Fu
tu
r
e
im
p
lem
en
tatio
n
s
co
u
l
d
f
u
r
th
e
r
ex
p
l
o
r
e
h
u
m
an
-
in
-
t
h
e
-
lo
o
p
a
p
p
r
o
ac
h
es,
allo
win
g
d
o
m
ain
ex
p
er
ts
to
v
alid
ate
m
o
d
el
o
u
tp
u
ts
an
d
p
r
o
v
id
e
co
r
r
ec
tiv
e
f
ee
d
b
ac
k
.
c.
B
r
o
ad
er
in
d
u
s
tr
ial
o
b
jectiv
es
B
ey
o
n
d
p
r
ed
ictiv
e
m
ain
ten
an
ce
an
d
p
r
o
d
u
ctio
n
q
u
ality
o
p
ti
m
izatio
n
,
f
u
tu
r
e
r
esear
ch
d
ir
e
ctio
n
s
m
ay
ex
ten
d
t
o
war
d
b
r
o
a
d
er
o
b
jecti
v
es
s
u
ch
as
e
n
er
g
y
ef
f
icien
cy
,
p
r
o
ce
s
s
s
u
s
tain
ab
ilit
y
,
a
n
d
en
v
ir
o
n
m
e
n
tal
im
p
ac
t
r
ed
u
ctio
n
.
T
h
ese
asp
ec
ts
alig
n
clo
s
ely
with
cu
r
r
en
t
in
d
u
s
tr
ial
p
r
io
r
ities
an
d
wo
u
ld
e
n
h
a
n
ce
th
e
s
o
cieta
l
an
d
o
p
er
atio
n
al
r
elev
an
ce
o
f
ML
-
d
r
iv
en
o
p
tim
izatio
n
s
y
s
tem
s
in
m
an
u
f
ac
tu
r
in
g
a
n
d
m
in
e
r
al
p
r
o
ce
s
s
in
g
s
ec
to
r
s
.
ACK
NO
WL
E
DG
M
E
N
T
S
T
h
e
au
th
o
r
s
wo
u
ld
lik
e
to
ex
p
r
ess
th
eir
s
in
ce
r
e
g
r
atitu
d
e
to
T
ah
r
i
Mo
h
am
m
e
d
Un
iv
er
s
ity
o
f
B
ec
h
ar
an
d
to
th
e
L
a
b
o
r
at
o
r
y
o
f
C
o
n
tr
o
l,
An
aly
s
is
,
an
d
Op
tim
izatio
n
o
f
E
lectr
o
-
E
n
er
g
etic
Sy
s
tem
s
(
C
AOSE
E
)
f
o
r
th
eir
in
s
titu
tio
n
al
an
d
tech
n
i
ca
l
s
u
p
p
o
r
t.
T
h
e
a
u
th
o
r
s
al
s
o
ac
k
n
o
wled
g
e
th
e
v
alu
ab
l
e
co
n
tr
ib
u
tio
n
s
o
f
co
lleag
u
es
an
d
ex
p
er
ts
wh
o
s
e
co
n
s
tr
u
ctiv
e
f
ee
d
b
ac
k
h
elp
ed
im
p
r
o
v
e
th
e
q
u
ality
o
f
t
h
is
r
esear
ch
.
Sp
ec
ial
th
an
k
s
ar
e
e
x
ten
d
e
d
to
t
h
e
n
at
io
n
al
in
itiativ
es
f
o
r
i
n
d
u
s
tr
ial
an
d
m
in
i
n
g
d
ev
elo
p
m
en
t
in
Alg
er
ia,
p
ar
ticu
lar
l
y
th
e
s
tr
ateg
ic
p
r
o
ject
f
o
r
th
e
e
x
p
lo
itatio
n
o
f
th
e
Gh
a
r
Djeb
ilet
ir
o
n
o
r
e
d
e
p
o
s
it,
wh
ich
in
s
p
i
r
ed
th
e
ca
s
e
s
tu
d
y
f
r
am
ewo
r
k
o
f
th
is
wo
r
k
.
Evaluation Warning : The document was created with Spire.PDF for Python.