I
AE
S
I
n
t
e
r
n
at
ion
al
Jou
r
n
al
of
Ar
t
if
icial
I
n
t
e
ll
ig
e
n
c
e
(
I
J
-
AI
)
Vol.
14,
No.
5:
Oc
tober
2025
,
pp.
4390
~
4402
I
S
S
N:
2252
-
8938
,
DO
I
:
10
.
11591/i
jai
.
v
14
.i
5
.
pp
43
90
-
4402
4390
Jou
r
n
al
h
omepage
:
ht
tp:
//
ij
ai
.
iaes
c
or
e
.
c
om
D
e
si
gn
in
g
an
in
t
e
ll
ig
e
n
t
sys
t
e
m
f
or
vi
b
r
at
io
n
d
ia
gn
osi
s
of
c
e
n
t
r
if
u
gal
w
at
e
r
-
c
ool
in
g
p
u
m
p
s
u
si
n
g
B
ay
e
si
an
n
e
t
w
o
r
k
s
Wiwit
S
u
p
r
ih
at
in
i
n
gs
ih
1
,
De
d
ik
Rom
ah
a
d
i
1
,
2
,
A
b
e
r
h
am
Genet
u
F
e
leke
2
1
D
e
pa
r
tm
e
nt
of
M
e
c
h
a
ni
c
a
l
E
ngi
ne
e
r
in
g,
F
a
c
ul
ty
of
E
ngi
ne
e
r
in
g,
U
ni
ve
r
s
it
a
s
M
e
r
c
u
B
ua
na
,
J
a
ka
r
ta
,
I
ndone
s
ia
2
S
c
hool
of
M
e
c
ha
ni
c
a
l
E
ngi
ne
e
r
in
g,
B
e
ij
in
g
I
ns
ti
tu
te
of
T
e
c
hno
lo
gy,
B
e
ij
in
g,
C
hi
na
Ar
t
icle
I
n
f
o
AB
S
T
RA
CT
A
r
ti
c
le
h
is
tor
y
:
R
e
c
e
ived
J
ul
4,
2024
R
e
vis
e
d
J
ul
27,
2025
Ac
c
e
pted
Aug
6,
2025
Imp
l
eme
n
t
i
n
g
mo
n
i
t
o
r
i
n
g
met
h
o
d
s
is
a
v
i
ab
l
e
met
h
o
d
to
red
u
ce
s
u
b
s
t
a
n
t
i
al
d
amag
e
to
c
o
o
l
i
n
g
w
a
t
er
cen
t
ri
f
u
g
a
l
p
u
m
p
s
.
E
n
g
a
g
i
n
g
in
ma
n
u
a
l
v
i
b
rat
i
o
n
an
al
y
s
i
s
req
u
i
re
s
co
n
s
i
d
erab
l
e
t
i
me
a
n
d
a
req
u
i
s
i
t
e
l
ev
el
of
co
m
p
et
e
n
ce.
Smal
l
d
at
a
s
et
s
p
o
s
e
c
h
al
l
en
g
es
w
h
e
n
ap
p
l
y
i
n
g
c
l
as
s
i
f
i
cat
i
o
n
s
y
s
t
ems
t
h
a
t
u
t
i
l
i
ze
l
i
n
ear
cl
a
s
s
i
fi
ca
t
i
o
n
m
o
d
e
l
s
an
d
d
ee
p
l
earn
i
n
g
.
G
i
v
en
t
h
e
s
e
i
s
s
u
es
,
our
p
ro
p
o
s
al
en
t
a
i
l
s
d
ev
e
l
o
p
i
n
g
a
s
y
s
t
em
cap
ab
l
e
of
au
t
o
n
o
mo
u
s
l
y
,
p
rec
i
s
e
l
y
,
an
d
accu
rat
el
y
d
i
ag
n
o
s
i
n
g
v
i
b
ra
t
i
o
n
s
u
s
i
n
g
a
l
i
mi
t
ed
d
at
as
e
t
.
The
s
y
s
t
em
is
an
t
i
ci
p
at
e
d
to
p
o
s
s
es
s
t
h
e
ca
p
ab
i
l
i
t
y
to
d
e
t
ect
mu
l
t
i
p
l
e
cat
e
g
o
r
i
es
of
mech
an
i
cal
d
efec
t
s
,
s
u
c
h
as
s
t
a
t
i
c
i
mb
a
l
a
n
ce,
d
y
n
am
i
c
i
mb
al
a
n
ce,
mi
s
a
l
i
g
n
me
n
t
,
ca
v
i
t
at
i
o
n
,
l
o
o
s
en
e
s
s
,
a
n
d
b
eari
n
g
co
rr
o
s
i
o
n
.
T
h
e
Bay
e
s
i
a
n
n
et
w
o
r
k
(BN
)
s
t
ru
c
t
u
re
w
as
co
n
s
t
ru
c
t
e
d
u
s
i
n
g
t
h
e
MA
T
L
A
B
s
o
f
t
w
are.
T
h
e
i
n
p
u
t
d
at
a
p
arame
t
ers
co
m
p
ri
s
e
v
i
b
rat
i
o
n
s
i
g
n
a
l
s
meas
u
red
in
t
h
e
freq
u
e
n
cy
d
o
ma
i
n
an
d
v
al
u
es
rep
r
es
e
n
t
i
n
g
p
h
a
s
e
d
i
ffere
n
ce
s
.
The
co
n
s
t
ru
c
t
ed
i
n
t
el
l
i
g
en
t
s
y
s
t
em
w
as
s
u
b
s
e
q
u
e
n
t
l
y
a
s
s
e
s
s
e
d
u
s
i
n
g
a
d
at
as
e
t
i
n
c
l
u
d
i
n
g
1
2
0
s
amp
l
es
.
T
h
e
s
mar
t
s
y
s
t
em
ca
n
rap
i
d
l
y
a
n
t
i
ci
p
at
e
a
n
d
p
rec
i
s
e
l
y
i
d
e
n
t
i
fy
ev
ery
fo
rm
of
h
arm
w
i
t
h
ex
cep
t
i
o
n
a
l
accu
racy
an
d
s
en
s
i
t
i
v
i
t
y
,
rel
y
i
n
g
on
t
es
t
o
u
t
co
me
s
.
The
t
e
s
t
d
a
t
a
an
al
y
s
i
s
rev
ea
l
s
t
h
a
t
the
i
n
t
e
l
l
i
g
en
t
s
y
s
t
em
at
t
ai
n
ed
an
av
erag
e
accu
rac
y
of
9
4
.
7
4
%
,
p
rec
i
s
i
o
n
of
9
5
.
3
2
%
,
s
en
s
i
t
i
v
i
t
y
(recal
l
)
of
9
3
.
6
7
%
,
an
d
F
-
s
c
o
re
of
9
4
.
3
6
%
.
K
e
y
w
o
r
d
s
:
B
a
ye
s
ian
ne
twor
ks
C
e
ntr
if
uga
l
pump
I
ntelli
ge
nt
s
ys
tem
M
a
c
hine
lea
r
ning
Vibr
a
ti
on
a
na
lys
is
Th
i
s
is
an
o
p
en
a
c
ces
s
a
r
t
i
c
l
e
u
n
d
e
r
the
CC
BY
-
SA
l
i
ce
n
s
e.
C
or
r
e
s
pon
din
g
A
u
th
or
:
De
dik
R
omaha
di
De
pa
r
tm
e
nt
of
M
e
c
ha
nica
l
E
nginee
r
ing
,
F
a
c
ult
y
of
E
nginee
r
ing
,
Unive
r
s
it
a
s
M
e
r
c
u
B
ua
na
J
a
ka
r
ta,
I
ndone
s
ia
E
mail:
de
dik
.
r
omaha
di
@me
r
c
ubua
na
.
a
c
.
id
1.
I
NT
RODU
C
T
I
ON
A
pump
is
a
de
v
ice
t
ha
t
c
a
n
mo
ve
f
lui
d
f
r
om
a
hi
ghe
r
e
leva
t
ion
or
an
a
r
e
a
wi
th
lo
we
r
pr
e
s
s
u
r
e
to
a
loca
ti
on
w
it
h
hi
ghe
r
p
r
e
s
s
ur
e
[
1
]
,
[
2
]
.
T
he
k
inet
i
c
e
ne
r
gy
or
ve
loc
it
y
of
t
he
i
mpel
ler
is
a
lt
e
r
e
d
d
ur
in
g
t
he
pumpi
ng
p
r
oc
e
s
s
to
f
a
c
il
i
tate
i
ts
r
ota
ti
on
a
nd
pr
ov
ide
the
ne
c
e
s
s
a
r
y
powe
r
f
o
r
f
l
uid
m
ove
me
nt
[
3]
.
Vibr
a
ti
o
n
r
e
s
ult
s
f
r
om
t
he
pu
mp's
r
otat
iona
l
mo
ve
men
t,
whic
h
c
a
us
e
s
the
i
mpel
ler
,
p
ump
s
ha
f
t
,
c
l
utch
,
a
n
d
oth
e
r
m
ovin
g
pa
r
ts
to
r
o
tate
.
E
xc
e
s
s
ive
vib
r
a
t
ion
in
the
p
ump
can
lea
d
to
de
t
r
i
menta
l
e
f
f
e
c
ts
s
uc
h
as
s
ha
f
t
a
nd
be
a
r
i
ng
da
mage
,
no
is
e
ge
ne
r
a
t
ion
,
de
c
r
e
a
s
e
d
he
a
d
a
n
d
c
a
pa
c
it
y
,
a
nd
e
ve
n
lo
we
r
pump
e
f
f
ic
ienc
y
[
4
]
.
Util
izing
pr
e
d
icti
ve
main
tena
nc
e
a
ppr
oa
c
he
s
can
he
l
p
take
p
r
e
ve
n
ti
ve
s
teps
to
li
m
it
th
e
li
ke
li
ho
od
of
a
br
u
pt
pump
br
e
a
k
d
owns
.
T
his
e
n
tails
pr
e
d
ict
ing
p
os
s
ibl
e
e
ngine
ha
r
m
by
e
xa
mi
ning
qua
n
ti
f
iab
le
pa
r
a
me
ter
s
a
nd
c
onduc
t
ing
a
c
o
mp
r
e
he
ns
i
ve
a
na
lys
is
.
Vi
br
a
ti
o
n
a
n
a
lys
is
is
a
p
r
oa
c
t
ive
mea
s
ur
e
to
mi
ti
ga
te
e
x
c
e
s
s
ive
vibr
a
ti
o
n
in
the
pum
p.
Vi
br
a
ti
on
can
a
d
ve
r
s
e
l
y
a
f
f
e
c
t
a
p
u
mp's
o
pe
r
a
ti
o
n,
lea
din
g
to
nois
e
,
r
e
du
c
e
d
pe
r
f
o
r
m
a
nc
e
,
a
nd
e
ve
n
ha
r
m
to
c
r
i
ti
c
a
l
c
ompo
ne
nts
l
ike
s
ha
f
ts
a
nd
be
a
r
in
gs
[
5
]
.
Vib
r
a
t
ion
a
na
lys
is
is
c
r
uc
ia
l
f
o
r
p
r
e
dicti
ng
da
mage
as
it
is
a
s
i
gni
f
ica
nt
t
ool
f
o
r
id
e
nt
if
y
ing
a
n
d
de
tec
t
ing
mec
ha
n
ica
l
is
s
ue
s
.
By
s
tudy
ing
c
ur
r
e
n
t
vibr
a
ti
o
n
c
ha
r
a
c
te
r
is
ti
c
s
,
it
is
pos
s
ib
le
to
e
va
luate
e
q
uip
ment
de
ter
io
r
a
ti
on
wit
hou
t
dis
a
s
s
e
mbl
y
[
6
]
,
[
7
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
Ar
ti
f
I
ntell
I
S
S
N:
2252
-
8938
De
s
igni
ng
an
int
e
ll
igent
s
y
s
tem
for
v
ibr
ati
on
diagn
os
is
of
c
e
ntr
if
ugal
w
ater
…
(
W
iw
it
Supr
ihat
ini
ngs
i
h
)
4391
T
he
mos
t
r
e
c
e
n
t
i
ter
a
ti
o
n
of
v
ib
r
a
ti
on
moni
to
r
in
g
e
qui
pmen
t
e
x
hib
it
s
e
n
ha
nc
e
d
c
a
pa
b
il
i
ti
e
s
a
n
d
a
utom
a
ted
ope
r
a
t
ions
c
om
pa
r
e
d
to
e
a
r
li
e
r
mo
de
ls
.
S
p
e
c
i
f
ic
de
v
ice
s
dis
pla
y
an
e
nt
ir
e
v
ibr
a
ti
o
n
s
pe
c
tr
u
m
on
thr
e
e
a
xe
s
c
onc
ur
r
e
nt
ly
,
p
r
ov
idi
n
g
a
ho
li
s
t
ic
de
p
ictio
n
of
the
ope
r
a
ti
o
na
l
c
ha
r
a
c
ter
is
ti
c
s
of
a
n
y
p
a
r
ti
c
ula
r
mac
hine
.
To
e
f
f
e
c
ti
v
e
ly
ut
il
ize
c
o
ntem
por
a
r
y
vi
b
r
a
ti
on
mea
s
u
r
ing
s
ys
tems
,
it
is
e
s
s
e
n
ti
a
l
to
ha
ve
a
ba
s
ic
unde
r
s
tandi
ng
of
vib
r
a
t
ion
a
na
lys
is
,
n
otw
it
hs
ta
n
ding
th
e
nume
r
ous
tec
hnol
ogica
l
f
e
a
tur
e
s
a
n
d
a
utom
a
ted
f
unc
ti
ons
they
of
f
e
r
.
Ana
l
yz
in
g
s
e
ve
r
a
l
pa
r
a
m
e
ter
s
in
c
o
ll
e
c
t
ing
vib
r
a
t
ion
da
ta
mi
g
ht
p
r
ov
ide
h
e
lp
f
ul
knowle
dge
of
the
e
xa
c
t
d
a
mage
inf
li
c
ted
on
the
e
q
uipm
e
nt
.
T
he
r
e
f
o
r
e
,
doi
ng
a
t
ho
r
oug
h
a
na
l
ys
is
can
ge
ne
r
a
te
s
uit
a
ble
r
e
c
om
menda
ti
ons
f
or
e
f
f
ic
ient
ly
ha
nd
li
n
g
e
ve
r
y
oc
c
ur
r
e
nc
e
of
e
quip
ment
f
a
il
ur
e
.
S
e
ve
r
a
l
vibr
a
ti
o
n
mea
s
ur
e
ments
ins
t
r
um
e
nts
do
not
pos
s
e
s
s
d
iagnos
t
ic
i
ntel
li
ge
nc
e
s
k
il
ls
,
pa
r
ti
c
u
lar
ly
whe
n
c
las
s
i
f
yi
ng
da
mag
e
c
a
us
e
d
by
t
he
im
ba
lanc
e
[
8]
,
[
9
]
.
Us
e
r
s
mus
t
ha
v
e
a
ba
s
ic
c
omp
r
e
he
ns
ion
of
the
f
u
nda
men
tals
of
vibr
a
ti
o
n
a
na
lys
is
c
onc
e
r
n
ing
d
if
f
e
r
e
nt
t
ype
s
of
s
t
r
uc
tu
r
a
l
da
mage
.
E
v
e
n
pr
of
e
s
s
i
ona
ls
r
e
qui
r
e
s
ig
ni
f
ica
n
t
t
im
e
to
diagnos
e
da
mage
c
a
us
e
d
by
im
ba
lanc
e
.
T
his
is
mainl
y
a
tt
r
ib
uted
to
the
c
o
ns
tr
a
ined
f
unc
ti
o
na
li
ti
e
s
of
the
e
xis
ti
n
g
t
ools
,
w
hich
ge
ne
r
a
l
ly
pr
o
vide
im
ba
lanc
e
d
da
ma
ge
pa
r
a
me
ter
s
.
T
he
r
e
f
or
e
,
the
us
e
r
mus
t
un
de
r
ta
ke
a
s
e
que
nc
e
of
da
ta
a
na
l
ys
is
meth
ods
to
a
s
c
e
r
tain
t
he
pr
e
c
is
e
na
tur
e
of
th
e
da
mage
t
ha
t
ha
s
take
n
plac
e
.
A
B
a
ye
s
ian
ne
twor
k
(
B
N)
is
a
model
that
is
r
e
p
r
e
s
e
nted
vis
ua
ll
y,
de
picting
p
r
oba
bil
is
ti
c
int
e
r
a
c
ti
ons
be
twe
e
n
int
e
r
c
onne
c
ted
va
r
iable
s
,
il
lus
tr
a
ti
ng
c
a
us
a
l
li
nka
ge
s
a
nd
de
pe
nde
nc
ies
[
10]
–
[
12]
.
BN
can
o
f
f
e
r
f
our
dis
ti
nc
t
methods
:
BN
de
mons
tr
a
tes
pr
o
f
icie
nc
y
in
pr
ope
r
ly
ha
ndli
ng
ins
tanc
e
s
of
da
ta
incomplet
e
ne
s
s
or
is
s
ue
s
.
M
or
e
ove
r
,
us
ing
BN
s
e
na
ble
s
the
ga
th
e
r
ing
of
knowle
dge
a
bout
c
a
us
e
-
a
nd
-
e
f
f
e
c
t
c
or
r
e
lations
.
Ga
ini
ng
inf
or
mation
a
nd
s
kil
ls
is
c
r
uc
ial
f
or
c
omp
r
e
he
nding
the
e
xtent
of
the
pr
oblem
domain
.
F
u
r
t
he
r
mor
e
,
BN
can
s
uppor
t
the
a
malga
mation
of
do
main
kno
wle
dge
a
nd
da
ta.
Ove
r
a
ll
,
BN
of
f
e
r
s
a
s
c
ientif
ica
ll
y
r
igor
ous
a
nd
e
f
f
icie
nt
a
ppr
oa
c
h
to
mi
ti
ga
te
the
is
s
ue
of
ove
r
f
it
ti
ng
in
da
ta
pr
oc
e
s
s
ing.
T
he
pr
oc
e
s
s
of
c
ons
tr
uc
ti
n
g
a
model
in
BN
c
ompr
is
e
s
two
s
e
pa
r
a
te
s
tage
s
:
th
e
f
ir
s
t
s
tep
include
s
e
s
tablis
hing
the
ne
twor
k
s
tr
uc
tur
e
.
In
c
ont
r
a
s
t,
the
s
ubs
e
que
nt
s
teps
e
ntail
e
s
ti
mating
the
pr
oba
bil
i
ty
va
lues
a
s
s
oc
iate
d
with
e
a
c
h
node
.
M
AT
L
AB
is
a
s
of
twa
r
e
pr
ogr
a
m
that
e
na
bles
th
e
c
ons
tr
uc
ti
on
of
BN
a
lgor
i
thm
s
.
Due
to
the
s
c
a
r
c
it
y
of
a
va
il
a
ble
da
tas
e
ts
,
the
nu
mer
ous
a
dva
ntage
s
of
BN
as
a
de
c
is
ion
-
making
tool
,
a
nd
the
e
ndo
r
s
e
ment
f
r
om
mul
ti
ple
s
c
ientif
ic
jou
r
na
ls
,
BN
wa
s
c
hos
e
n
as
the
pr
e
f
e
r
r
e
d
method
f
or
c
r
e
a
ti
ng
a
diagnos
ti
c
s
y
s
tem
f
or
wa
ter
-
c
ooled
c
e
ntr
if
uga
l
pump
e
ngines
[
13
]
,
[
14]
.
N
u
m
e
r
ou
s
i
n
qu
i
r
i
e
s
h
a
ve
b
e
e
n
c
on
d
uc
t
e
d
in
t
h
is
a
r
e
a
of
r
e
s
e
a
r
c
h
[
1
5
]
–
[1
7]
.
O
u
r
p
r
e
v
io
us
s
t
u
dy
a
l
s
o
d
i
s
c
us
s
e
d
t
he
de
s
ig
n
of
an
i
nt
e
l
l
ig
e
nt
s
y
s
te
m
to
de
t
e
c
t
t
y
pe
s
of
i
m
ba
l
a
n
c
e
s
by
u
t
il
i
z
in
g
B
N
,
a
nd
s
a
ti
s
f
a
c
t
o
r
y
r
e
s
u
lt
s
w
e
r
e
o
b
t
a
i
n
e
d
[
1
8
]
.
T
h
e
da
m
a
g
e
l
oc
a
ti
o
n
can
be
de
t
e
r
m
in
e
d
by
e
xa
m
i
ni
n
g
t
he
v
ib
r
a
t
i
on
s
ig
n
a
l
,
in
w
h
ic
h
t
h
e
a
m
p
l
i
tu
d
e
of
t
h
e
v
i
b
r
a
t
i
o
n
w
i
l
l
be
q
ua
n
t
i
f
ie
d
.
T
h
e
r
e
s
u
l
ts
w
i
l
l
be
d
is
p
la
y
e
d
in
t
h
e
f
r
e
q
ue
nc
y
d
o
m
a
in
u
t
i
l
i
z
in
g
t
h
e
f
a
s
t
F
o
u
r
i
e
r
t
r
a
ns
f
o
r
m
(
F
F
T
)
a
p
p
r
o
a
c
h
.
T
h
e
s
t
ud
y
c
o
n
du
c
te
d
by
C
a
s
t
e
l
la
n
os
et
al
.
[
3
]
s
p
e
c
i
f
i
c
a
l
l
y
e
x
a
m
in
e
d
t
h
e
a
n
a
ly
s
is
of
v
i
b
r
a
t
i
o
ns
in
m
o
to
r
s
a
n
d
p
u
m
ps
.
T
he
v
i
b
r
a
t
i
o
n
a
na
l
ys
is
wa
s
p
e
r
f
o
r
me
d
ma
n
ua
l
l
y
u
s
in
g
t
h
e
FFT
t
e
c
hn
o
l
o
gy
.
T
h
i
s
m
e
t
h
o
d
is
e
mp
l
o
ye
d
to
a
s
s
e
s
s
th
e
e
x
te
n
t
of
ha
r
m
i
n
f
l
i
c
t
e
d
on
t
h
e
p
um
p
,
as
e
v
i
de
n
c
e
d
by
f
r
e
q
u
e
nc
y
s
p
e
c
t
r
u
m
me
a
s
u
r
e
m
e
n
ts
.
E
m
p
i
r
ic
a
l
da
t
a
in
d
i
c
a
t
e
s
t
ha
t
h
a
r
mo
n
i
c
a
m
p
l
it
u
d
e
s
me
a
s
u
r
e
d
at
wh
o
le
n
u
m
be
r
m
u
l
t
i
p
le
s
of
e
n
gi
n
e
s
pe
e
d
,
p
r
e
c
is
e
l
y
at
1,
2,
a
n
d
3,
c
a
n
be
u
s
e
d
as
a
d
e
pe
n
da
b
le
s
i
g
n
a
l
f
o
r
d
e
te
c
t
i
n
g
m
is
a
l
i
g
n
m
e
n
t
i
s
s
u
e
s
in
t
h
e
p
r
e
s
e
nc
e
of
v
a
r
ia
t
i
o
ns
in
m
o
t
o
r
l
e
g
he
i
gh
t
.
In
a
d
d
i
t
i
o
n
,
r
e
s
e
a
r
c
h
e
r
s
h
a
ve
e
x
a
m
in
e
d
c
e
n
t
r
i
f
uga
l
p
u
m
ps
u
t
i
l
i
z
e
d
in
c
o
o
l
in
g
w
a
te
r
s
ys
t
e
ms
to
a
s
s
e
s
s
t
h
e
m
a
g
n
it
u
d
e
of
v
i
b
r
a
t
io
n
-
i
n
du
c
e
d
h
a
r
m
.
T
h
e
p
u
m
p
's
v
i
b
r
a
ti
on
d
a
t
a
is
s
u
bs
e
q
u
e
n
t
l
y
p
r
o
c
e
s
s
e
d
u
s
in
g
FFT
t
e
c
h
n
i
qu
e
s
to
e
n
a
b
l
e
v
i
s
u
a
l
e
xa
m
i
na
t
i
o
n
a
nd
d
e
t
e
r
mi
n
a
t
io
n
of
t
h
e
s
p
e
c
i
f
i
c
t
y
p
e
of
d
a
m
a
g
e
th
a
t
ha
s
o
c
c
u
r
r
e
d
[
1
9
]
.
A
na
l
ys
is
of
t
he
v
i
b
r
a
t
i
o
n
s
pe
c
t
r
um
i
n
d
ic
a
tes
t
h
e
p
r
e
s
e
nc
e
of
me
c
ha
n
i
c
a
l
l
o
o
s
e
n
e
s
s
d
a
m
a
ge
in
t
he
m
o
t
o
r
a
n
d
p
u
mp
c
o
mp
on
e
n
ts
.
T
h
e
e
ng
i
ne
c
o
nd
i
t
i
o
n
a
s
s
e
s
s
me
n
t
is
o
f
t
e
n
c
o
n
d
u
c
t
e
d
a
c
c
o
r
d
i
n
g
to
th
e
I
S
O
1
08
1
6
-
3
s
ta
n
da
r
d
,
e
nc
o
mp
a
s
s
in
g
s
p
e
e
d
a
n
d
a
c
c
e
le
r
a
t
i
on
m
od
e
s
.
A
c
c
o
r
d
in
g
to
d
a
t
a
s
p
e
c
t
r
u
m
i
n
v
e
s
t
i
ga
t
i
on
s
,
t
he
p
r
i
m
a
r
y
c
a
u
s
e
of
d
a
m
a
ge
is
r
o
t
o
r
i
m
b
a
l
a
nc
e
.
W
i
th
i
n
v
ib
r
a
t
io
n
a
na
l
ys
is
,
e
x
pe
r
t
s
ys
t
e
ms
p
r
i
o
r
i
t
i
z
e
t
h
e
de
pe
n
d
a
b
i
l
i
ty
of
BN
s
in
e
f
f
ic
i
e
n
t
l
y
a
d
d
r
e
s
s
i
n
g
c
i
r
c
u
ms
ta
n
c
e
s
t
h
a
t
in
v
o
l
ve
u
nc
e
r
ta
i
n
t
y
.
T
h
e
BN
m
e
t
h
o
d
is
w
id
e
l
y
r
e
g
a
r
de
d
as
a
p
r
a
c
t
ic
a
l
wa
y
to
de
v
e
l
o
p
i
n
t
e
l
l
i
g
e
n
t
v
i
b
r
a
t
i
o
n
de
t
e
c
t
i
on
a
p
p
li
c
a
t
i
o
ns
[
2
0
]
–
[
2
2
]
.
O
u
r
p
r
o
p
os
e
d
r
e
s
e
a
r
c
h
a
i
ms
to
p
r
o
v
i
de
a
n
o
v
e
l
a
p
p
r
o
a
c
h
f
o
r
c
o
n
s
t
r
uc
t
i
ng
a
s
ys
t
e
m
c
a
p
a
b
l
e
of
d
i
a
g
no
s
in
g
da
m
a
g
e
to
t
he
P
-
9
11
4
B
t
y
pe
wa
t
e
r
c
o
o
li
n
g
c
e
n
t
r
i
f
u
ga
l
pu
m
p
us
in
g
v
ib
r
a
t
io
n
da
t
a
.
We
ha
ve
de
v
e
lo
p
e
d
a
m
e
th
o
d
to
i
n
te
r
p
r
e
t
t
h
e
r
oo
t
me
a
n
s
qu
a
r
e
(
R
M
S
)
as
a
p
e
r
f
o
r
ma
nc
e
in
d
i
c
a
t
o
r
ba
s
e
d
on
t
he
I
S
O
1
0
8
16
-
3
s
t
a
n
da
r
d
.
We
u
t
il
i
z
e
a
no
v
e
l
BN
mo
d
e
l
to
a
n
a
ly
z
e
t
h
e
s
pe
c
t
r
um
,
e
n
a
b
li
n
g
it
to
m
a
k
e
de
t
e
r
mi
n
a
t
io
ns
b
a
s
e
d
on
th
e
s
p
e
c
i
f
ic
t
y
p
e
of
m
a
c
h
in
e
r
y
d
a
m
a
ge
.
T
h
i
s
s
t
u
dy
y
i
e
l
de
d
a
h
i
g
h
l
y
e
f
f
ic
i
e
n
t
a
p
p
r
o
a
c
h
t
h
a
t
can
be
u
t
i
l
iz
e
d
e
x
p
li
c
i
t
l
y
f
o
r
d
i
a
g
no
s
in
g
p
u
mp
s
.
T
h
is
r
e
s
e
a
r
c
h
a
i
ms
to
a
i
d
i
n
du
s
tr
y
p
e
r
s
o
n
ne
l
in
e
f
f
ic
i
e
n
t
l
y
a
n
d
a
c
c
u
r
a
t
e
ly
i
d
e
n
t
i
f
y
ing
v
a
r
i
o
us
t
y
p
e
s
of
m
a
c
h
i
ne
d
a
ma
ge
,
e
v
e
n
w
i
t
ho
u
t
a
c
o
m
p
r
e
he
n
s
i
v
e
kn
o
w
le
d
ge
of
t
h
e
f
un
d
a
m
e
n
ta
l
p
r
i
nc
i
p
le
s
a
nd
i
d
e
a
s
of
v
i
b
r
a
t
io
n
a
n
a
l
y
s
i
s
.
R
e
s
e
a
r
c
h
i
n
g
t
he
u
t
i
l
i
z
a
t
i
on
of
BN
to
d
i
a
g
n
os
e
i
n
d
us
t
r
ia
l
m
a
c
h
i
n
e
r
y
.
T
h
e
r
e
s
e
a
r
c
h
f
i
n
di
n
g
s
c
a
n
p
o
t
e
n
ti
a
l
ly
s
e
r
v
e
as
a
ba
s
is
f
o
r
e
s
t
a
b
l
is
h
i
ng
a
d
i
a
g
n
os
t
i
c
s
ys
te
m
on
a
b
r
oa
de
r
s
c
a
l
e
.
2.
M
E
T
HO
D
P
u
mp
v
ib
r
a
t
io
n
da
ta
wa
s
o
bt
a
i
ne
d
t
h
r
o
ug
h
th
e
c
o
n
duc
t
io
n
of
v
ib
r
a
t
io
n
mea
s
u
r
e
me
nts
.
Da
ta
wa
s
c
ol
le
c
t
e
d
us
in
g
t
he
V
ibX
pe
r
t
II
ins
t
r
u
me
nt
at
b
ot
h
t
he
b
e
a
r
i
ngs
at
t
he
d
r
ive
e
n
d
(
DE
)
a
n
d
th
e
o
ne
at
th
e
no
n
-
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14,
No.
5:
Oc
tober
2025
:
43
90
-
4402
4392
d
r
iv
e
e
n
d
(
ND
E
)
p
os
i
t
io
ns
.
M
e
a
s
ur
e
m
e
n
ts
we
r
e
c
on
d
uc
t
e
d
in
h
o
r
i
z
o
nta
l
,
v
e
r
t
ica
l
,
a
n
d
a
x
ia
l
o
r
ien
ta
t
io
ns
on
bo
th
t
he
m
ot
or
a
n
d
p
um
p
s
ides
.
A
f
t
e
r
t
ha
t
,
t
he
d
a
t
a
p
r
o
c
e
s
s
in
g
o
pe
r
a
ti
on
wa
s
e
xe
c
u
ted
us
in
g
M
A
T
L
A
B
s
o
f
tw
a
r
e
.
2.
1.
Vib
r
at
io
n
d
a
t
a
ac
q
u
is
it
ion
Vibxpe
r
t
II
mea
s
ur
e
ment
ins
tr
uments
a
r
e
e
mpl
oye
d
in
thi
s
inves
ti
ga
ti
on
to
qua
nti
f
y
pump
vib
r
a
ti
ons
a
s
s
hown
in
F
igur
e
1
.
T
he
ins
tr
uments
us
e
d
to
a
c
quir
e
vibr
a
ti
ona
l
da
ta
a
r
e
de
picte
d
in
F
igu
r
e
1(
a
)
.
T
a
ble
1
dis
plays
the
s
pe
c
if
ica
ti
ons
of
the
Vibxpe
r
t
II
vibr
a
ti
on
mea
s
ur
ing
de
vice
,
while
T
a
ble
2
dis
plays
the
s
pe
c
if
ica
ti
ons
of
the
v
ibr
a
ti
on
a
c
c
e
ler
omete
r
s
e
ns
or
,
whic
h
a
r
e
6
.
142
.
F
or
thi
s
a
na
lys
is
,
we
uti
li
z
e
d
a
c
ooli
ng
wa
ter
c
e
ntr
if
uga
l
pump,
de
picte
d
in
F
igur
e
1(
b
)
,
a
c
r
uc
ial
e
leme
nt
of
the
c
ooli
ng
towe
r
s
ys
t
e
m.
T
he
c
e
ntr
if
uga
l
pump,
r
otating
at
a
s
pe
e
d
of
1
,
450
r
p
m
,
ha
s
a
powe
r
output
of
114.
4
kW
.
At
PT
S
AU
,
th
e
f
lui
d
is
tr
a
ns
f
e
r
r
e
d
f
r
om
a
c
ooli
ng
towe
r
to
a
he
a
t
e
xc
ha
ng
e
r
us
ing
the
P
9114B
pump
.
T
he
P
9114B
DE
pump
uti
li
z
e
s
a
3314
-
type
a
ngular
c
ontac
t
be
a
r
ing,
while
the
N
DE
a
nd
DE
-
NDE
mot
or
s
ide
uti
li
z
e
s
a
ba
ll
c
us
hion
in
the
6314
-
type
gr
oo
ve
.
T
he
g
land
pa
c
king
kind
is
us
e
d
to
s
e
a
l
a
s
be
s
tos
r
a
ther
than
pr
ovide
lubr
ica
ti
on.
T
he
s
pe
c
s
f
or
the
P
9114B
c
e
ntr
if
uga
l
c
ooli
ng
wa
ter
pump
a
r
e
dis
playe
d
in
T
a
ble
3.
(
a
)
(
b)
F
igur
e
1
.
Vibr
a
ti
on
mea
s
ur
e
ment
s
e
tup
of
c
e
ntr
if
u
ga
l
pump
P
9114
B
,
(
a
)
VibXpe
r
t
II
da
ta
c
oll
e
c
tor
with
s
e
ns
or
6.
142
us
e
d
f
or
vibr
a
ti
on
a
na
lys
is
a
nd
(
b)
c
e
ntr
if
uga
l
pump
with
labe
led
mea
s
ur
e
ment
point
s
at
DE
-
NDE
mot
or
a
nd
pump
T
a
ble
1
.
VibXpe
r
t
II
s
pe
c
if
ica
ti
on
N
a
me
S
pe
c
if
ic
a
ti
on
R
a
nge
f
r
e
que
nc
y
0.5
Hz
–
40
kH
z
E
nvi
r
onme
nt
pr
ot
e
c
ti
on
I
P
65
T
e
mpe
r
a
tu
r
e
ope
r
a
ti
ona
l
0
-
50
°
C
D
a
ta
me
mor
i
128
MB
DDR
R
A
M
T
ip
e
ba
te
r
a
i
Li
-
io
n
r
e
c
har
ge
abl
e
(
7.3
V
/5
.3
Ah
–
38.7
W
h)
D
im
e
ns
io
n
186
×
162
×
52
mm
(
L
xW
xH
)
T
a
ble
2
.
Vibr
a
ti
on
s
e
ns
or
s
pe
c
if
ica
ti
on
N
a
me
S
pe
c
if
ic
a
ti
on
T
r
a
ns
mi
s
io
n
f
a
c
to
r
1
.
0
µA/m
/s
2
R
e
s
ona
nc
e
f
r
e
que
nc
y
36
kH
z
T
e
mpe
r
a
tu
r
e
r
a
nge
-
40
-
100
°
C
C
a
s
e
ma
t
e
r
ia
l
S
ta
in
le
s
s
s
te
e
l
VA
1.4305
E
nvi
r
onme
nt
pr
ot
e
c
ti
on
IP
65
T
a
ble
3
.
C
e
ntr
if
uga
l
p
u
mp
P
9114B
s
pe
c
if
ica
ti
on
N
a
me
S
pe
c
if
ic
a
ti
on
P
ump
ty
pe
C
e
nt
r
if
uga
l
pump
M
ode
l
3K
10
X
8
-
16/
156
B
r
a
nd
D
U
R
C
O
M
A
R
K
III
S
uc
ti
on
pr
e
s
s
ur
e
5.5
B
a
r
R
P
M
1
,
450
P
ow
e
r
114.4
kW
2.
2.
P
r
i
m
ar
y
f
a
u
lt
f
r
e
q
u
e
n
c
y
c
om
p
on
e
n
t
F
r
e
que
nc
y
r
e
quir
e
ments
a
nd
c
omput
a
ti
ons
a
r
e
c
r
u
c
ial
in
inves
ti
ga
ti
ng
mac
hine
f
a
il
ur
e
,
pa
r
t
icula
r
ly
be
a
r
ing
f
r
e
que
nc
ies
[
23]
.
T
he
pump
uti
li
z
e
s
type
6
,
314
ba
ll
be
a
r
ings
a
nd
is
a
c
c
ompanie
d
by
c
ompr
e
he
ns
ive
s
pe
c
if
ica
t
ions
.
In
(
1
)
to
(
4
)
dis
play
the
f
r
e
que
nc
ies
s
pe
c
if
ic
to
d
if
f
e
r
e
nt
be
a
r
ing
c
omponents
,
s
uc
h
as
t
he
inner
a
nd
outer
r
a
c
e
s
,
ba
ll
s
or
r
oll
e
r
s
,
a
nd
c
a
ge
s
.
If
a
be
a
r
ing
e
leme
nt
f
r
e
que
nc
y
is
c
onf
ir
med,
it
s
igni
f
ies
a
be
a
r
ing
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
Ar
ti
f
I
ntell
I
S
S
N:
2252
-
8938
De
s
igni
ng
an
int
e
ll
igent
s
y
s
tem
for
v
ibr
ati
on
diagn
os
is
of
c
e
ntr
if
ugal
w
ater
…
(
W
iw
it
Supr
ihat
ini
ngs
i
h
)
4393
is
s
ue
a
s
s
oc
iate
d
with
pump
da
mage
.
In
(
1
)
can
be
us
e
d
to
c
a
lcula
te
the
f
r
e
que
nc
y
of
ba
ll
pa
s
s
e
s
f
r
e
que
nc
y
in
the
inner
r
ing
(
B
P
F
I
)
.
=
120
(
1
+
)
(
1
)
W
he
r
e
B
D
is
a
ba
ll
diame
ter
of
20
mm
,
P
D
is
a
pit
c
h
diame
ter
of
100
mm,
In
thi
s
c
a
s
e
,
the
c
ontac
t
a
ngle
is
0
o
,
a
nd
the
mo
tor
c
e
ntr
i
f
uga
l
pump
s
pe
e
d
is
1
,
450
r
pm
or
24
.
17
Hz
.
In
(
2
)
can
be
us
e
d
to
de
ter
mi
ne
the
outer
r
ing's
ba
ll
pa
s
s
f
r
e
que
nc
y
o
uter
(
B
P
F
O)
.
=
120
(
1
+
)
⋅
(
2
)
W
he
r
e
B
N
is
the
numbe
r
of
ba
ll
be
a
r
ings
e
qua
l
to
8
ba
ll
be
a
r
ings
.
F
unda
menta
l
tr
a
in
f
r
e
que
nc
y
(
F
T
F
)
is
c
a
lcula
ted
us
ing
(
3
)
.
=
120
(
1
−
)
(
3
)
M
e
a
nwhile,
ba
ll
s
pin
f
r
e
que
nc
y
(
B
S
F
)
can
be
obtai
ne
d
f
r
om
(
4
)
.
=
120
(
1
−
(
)
2
)
⋅
(
4
)
T
he
c
e
ntr
i
f
uga
l
pump
im
pe
ll
e
r
ha
s
s
ix
blade
a
ngle
s
.
Us
e
(
5
)
to
c
omput
e
the
f
r
e
que
nc
y
ge
ne
r
a
ted
by
t
he
pump
im
pe
ll
e
r
,
i.
e
.
,
ba
ll
pa
s
s
f
r
e
que
nc
y
(
B
P
F
)
.
=
60
×
(
5
)
In
a
ddit
ion
to
the
f
r
e
que
nc
ies
mentioned
in
(
1
)
to
(
5
)
,
a
nother
s
igni
f
ica
nt
f
r
e
que
nc
y
is
the
one
a
s
s
oc
iate
d
with
mot
or
s
pinni
ng
.
T
he
mot
o
r
r
otation
f
r
e
que
nc
y
of
21.
17
Hz
is
c
omm
only
r
e
f
e
r
r
e
d
to
as
1X
a
nd
can
a
ls
o
be
r
e
f
e
r
r
e
d
to
as
mul
ti
ples
,
s
uc
h
as
2X,
3X
,
a
nd
so
f
or
th.
2.
3.
Vib
r
at
io
n
a
s
s
e
s
s
m
e
n
t
T
he
time
domain
vibr
a
ti
on
da
ta
is
obtaine
d
us
ing
t
he
R
M
S
a
ppr
oa
c
h,
as
in
(
6
)
[
24
]
.
T
he
f
ind
ings
of
the
R
M
S
c
omput
a
ti
on
will
be
co
mpar
e
d
a
ga
ins
t
the
I
S
O
10816
-
3
s
tanda
r
d.
T
his
s
tudy
a
im
s
to
a
s
s
e
s
s
the
pump's
magnitude
a
nd
s
tate
of
vibr
a
ti
on
.
I
S
O
108
16,
as
r
e
f
e
r
e
nc
e
d
in
[
25]
,
outl
ines
s
tanda
r
ds
f
or
e
v
a
luating
whe
n
taking
r
e
a
dings
of
vibr
a
ti
on
leve
ls
dir
e
c
tl
y
at
the
mea
s
ur
e
ment
s
it
e
.
=
√
1
∑
|
|
2
=
1
(
6
)
T
h
e
a
bo
ve
c
r
i
te
r
ia
pe
r
ta
in
to
mac
hi
ne
de
v
ice
s
w
it
h
a
p
owe
r
r
a
t
i
ng
e
x
c
e
e
di
ng
15
kW
a
n
d
an
op
e
r
a
t
io
na
l
s
pe
e
d
r
a
n
gi
ng
f
r
o
m
12
0
to
15
,
00
0
r
e
vo
lu
ti
ons
pe
r
m
i
nu
te
.
T
he
I
S
O
1
08
16
-
3
:
20
09
s
tan
da
r
d
,
ba
s
e
d
on
v
e
l
oc
it
y
m
e
a
s
u
r
e
men
ts
,
e
v
a
l
ua
tes
t
he
v
ib
r
a
t
io
n
le
ve
ls
in
m
a
c
hi
ne
s
w
it
h
a
r
ot
a
t
in
g
s
pe
e
d
e
x
c
e
e
di
ng
6
00
r
p
m
.
V
i
br
a
t
ion
a
s
s
e
s
s
m
e
n
t
t
yp
ic
a
l
ly
us
e
s
un
it
s
of
m
m/
s
R
M
S
.
S
u
pp
os
e
th
e
R
M
S
va
lu
e
is
un
de
r
t
he
s
pe
c
i
f
i
e
d
li
m
it
of
4
.
5
mm
/s
,
as
s
ta
ted
in
t
he
I
S
O
10
81
6
-
3
g
r
ou
p
2
s
ta
nd
a
r
d
f
o
r
a
r
i
gi
d
f
ou
nda
t
io
n
ty
pe
.
In
t
ha
t
c
a
s
e
,
th
e
vi
b
r
a
ti
on
l
e
v
e
l
of
the
p
u
mp
is
c
ons
id
e
r
e
d
a
c
c
e
pta
b
le
.
T
he
r
e
f
o
r
e
,
th
e
r
e
is
no
ne
e
d
to
c
on
t
in
ue
wi
t
h
th
e
s
ubs
e
qu
e
n
t
p
r
o
c
e
du
r
e
.
H
ow
e
ve
r
,
if
t
he
vi
b
r
a
ti
on
s
u
r
pa
s
s
e
s
4
.
5
m
m
/s
,
it
is
ne
c
e
s
s
a
r
y
to
e
xa
mi
ne
a
nd
id
e
n
ti
f
y
t
he
or
i
gi
n
of
the
da
m
a
g
e
.
2.
4
.
S
ys
t
e
m
f
lowc
h
ar
t
F
igur
e
2
de
picts
a
de
tailed
f
lowc
ha
r
t
il
lus
tr
a
ti
ng
the
f
low
of
a
pa
r
t
icula
r
s
ys
tem.
T
he
BN
method
will
c
a
lcula
te
the
pr
oba
bil
it
y
va
lue
us
ing
the
in
put
da
ta.
T
his
pr
oc
e
s
s
invol
ve
s
s
e
ve
r
a
l
s
tage
s
,
i
nc
ludi
ng
s
olvi
ng
f
or
each
da
mage
s
ympt
om
by
f
indi
ng
its
p
a
r
a
mete
r
va
lue
a
nd
c
a
lcula
ti
ng
the
pr
oba
bil
it
y
va
l
ue
unde
r
c
e
r
tain
c
ondit
ions
.
Onc
e
the
da
ta
is
c
oll
e
c
ted,
t
he
a
lgor
it
hm
wil
l
de
te
r
mi
ne
the
c
ombi
ne
d
a
nd
pos
ter
ior
pr
oba
bil
it
y
va
lue
f
or
e
ve
r
y
f
or
m
of
da
mage
.
T
he
BN
f
r
a
mew
or
k
is
c
ha
nge
d,
a
nd
the
pos
ter
ior
pr
oba
bil
it
y
e
s
ti
mation
is
ut
il
ize
d
to
make
a
pr
oba
bil
it
y
inf
e
r
e
n
c
e
r
e
ga
r
ding
the
da
mage
.
BN
ut
il
ize
s
two
-
wa
y
pr
o
pa
ga
ti
on
be
twe
e
n
node
s
that
r
e
c
e
ive
a
nd
s
e
nd
da
ta
to
pr
ovide
r
e
lational
inf
or
mation
a
nd
c
ondit
ional
pr
ob
a
bil
it
ies
.
In
pr
a
c
ti
c
a
l
a
ppli
c
a
ti
ons
,
mul
ti
s
tate
node
s
a
r
e
c
o
mm
only
us
e
d.
B
a
s
e
d
on
the
us
e
r
's
c
ons
ult
a
ti
on,
t
he
y
will
r
e
c
e
ive
inf
or
mation
on
the
s
pe
c
if
ic
da
mage
to
the
mac
hine
a
nd
the
pe
r
c
e
ntage
of
e
r
r
o
r
s
.
Us
e
r
s
a
r
e
r
e
quir
e
d
to
c
ompi
le
da
ta
in
the
f
o
r
mat
of
mac
hine
s
pe
c
s
a
nd
vibr
a
ti
on
da
ta.
T
he
s
ys
tem
will
a
na
lyze
the
number
s
that
make
up
the
vibr
a
ti
on
s
pe
c
tr
um
to
ge
ne
r
a
te
in
f
or
mation
whos
e
outpu
t
the
BN
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14,
No.
5:
Oc
tober
2025
:
43
90
-
4402
4394
can
uti
li
z
e
.
T
he
pr
ovided
s
pe
c
t
r
um
da
ta
is
f
u
r
th
e
r
a
na
lyze
d
to
de
ter
mi
ne
the
f
r
e
que
nc
y
of
each
mac
hine
c
omponent.
T
he
de
tails
a
bout
the
mac
hine's
s
pe
c
if
ica
ti
ons
will
be
a
r
e
f
e
r
e
nc
e
f
or
f
r
e
que
nc
y
c
a
lcula
ti
ons
a
nd
will
be
f
ine
-
tuned
to
matc
h
the
s
pe
c
tr
um.
If
the
c
omput
e
d
f
r
e
que
nc
y
va
lue
is
pr
e
s
e
nt
in
the
s
pe
c
tr
um,
it
will
s
e
r
ve
as
e
videnc
e
f
or
B
N.
C
onve
r
s
e
ly,
if
the
f
r
e
qu
e
nc
y
va
lue
is
not
f
ound
in
the
s
pe
c
tr
um,
it
will
not
pr
ovide
a
ny
pr
oof
of
J
B
.
J
B
's
input
da
ta
is
c
a
tegor
ize
d
int
o
two
dis
ti
nc
t
kinds
.
F
r
e
que
nc
y
li
ne
s
e
mpi
r
ica
ll
y
ve
r
if
ied
by
f
r
e
que
nc
y
c
a
lcula
ti
on
a
r
e
c
las
s
if
ied
as
input
type
1
(
tr
ue
)
.
T
his
input
r
e
quir
e
s
that
the
f
r
e
que
nc
y
a
mpl
it
ude
be
e
qua
l
to
or
gr
e
a
ter
than
2
mm
/s
R
M
S
.
How
e
ve
r
,
if
the
da
ta
doe
s
not
c
ontain
a
f
r
e
que
nc
y
li
ne
s
uppor
te
d
by
the
c
omput
a
ti
on
or
a
f
r
e
que
nc
y
a
mpl
i
tude
va
lue
of
les
s
than
2
mm/s
R
M
S
,
it
is
c
las
s
if
ied
as
input
type
2
(
f
a
ls
e
)
.
O
ve
r
a
l
l
V
i
br
a
t
i
on V
a
l
ue
V
a
l
ue
S
e
t
E
vi
de
nc
e
1
E
vi
de
nc
e
2
B
a
ye
s
i
a
n N
e
t
w
or
ks
C
om
pu
t
a
t
i
on
I
nf
e
r
e
nc
e
P
r
oba
bi
l
i
t
y
D
i
a
gnos
i
s
R
e
s
ul
t
R
e
s
e
t
F
i
ni
s
h
S
t
a
r
t
-
M
a
c
hi
ne
S
pe
c
i
f
i
c
a
t
i
on
-
V
i
br
a
t
i
on
D
a
t
a
C
or
r
e
c
t
No
S
pe
c
t
r
um
L
i
ne
I
de
nt
i
f
i
c
a
t
i
on
Y
e
s
No
No
A
A
Y
es
F
igur
e
2
.
S
ys
tem
f
lowc
ha
r
t
2.
5
.
S
ys
t
e
m
p
e
r
f
or
m
an
c
e
c
alcul
a
t
ion
s
As
s
e
s
s
ment
is
c
onduc
ted
withi
n
the
f
r
a
mew
or
k
of
s
upe
r
vis
e
d
lea
r
ning.
As
the
li
te
r
a
tur
e
il
lus
tr
a
tes
,
each
matr
ix
r
ow
de
notes
an
a
c
tual
c
las
s
oc
c
ur
r
e
nc
e
,
while
each
c
olum
n
de
notes
a
pr
e
dicte
d
c
las
s
oc
c
ur
r
e
nc
e
.
An
a
lt
e
r
na
ti
ve
s
tr
uc
tur
e
pr
opos
e
d
in
the
a
c
a
de
mi
c
li
te
r
a
tur
e
e
ntails
the
a
li
gnment
of
each
r
ow
with
a
nti
c
ipate
d
c
las
s
e
ve
nts
a
nd
each
c
olum
n
with
a
c
tual
c
las
s
e
ve
nts
.
T
his
nomenc
latur
e
a
im
s
to
e
nh
a
nc
e
the
identif
ica
ti
on
p
r
oc
e
s
s
if
the
s
ys
tem
inac
c
ur
a
tely
a
s
s
igns
the
s
a
me
c
a
tegor
y
to
two
dis
ti
nc
t
gr
oups
,
f
r
e
que
ntl
y
lea
ding
to
mi
s
c
las
s
if
ica
ti
on.
T
he
e
xa
mi
ne
d
c
onti
nge
nc
y
table
c
ompr
is
e
s
two
dim
e
n
s
ions
:
"
a
c
tual"
a
nd
"
pr
e
dicte
d.
"
B
oth
dim
e
ns
ions
e
nc
ompas
s
identica
l
"
c
las
s
e
s
.
"
E
a
c
h
va
r
iable
in
a
c
ont
ingenc
y
ta
ble
is
a
dis
ti
nc
ti
ve
c
ombi
na
ti
on
of
c
las
s
if
ica
ti
ons
a
nd
di
mens
i
ons
.
A
t
r
ue
pos
it
ive
(
T
P
)
is
a
s
it
ua
ti
on
in
whic
h
the
a
c
tual
va
lue
a
nd
the
pr
e
diction
a
r
e
both
pos
it
ive.
A
tr
ue
ne
ga
ti
ve
(
T
N
)
is
a
s
it
ua
ti
on
in
whic
h
both
t
he
pr
e
c
is
e
number
a
nd
the
pr
e
diction
a
r
e
ne
ga
ti
ve
.
A
f
a
ls
e
pos
it
ive
(
F
P
)
is
a
s
it
ua
ti
on
in
whic
h
the
p
r
e
diction
is
pos
it
ive,
e
ve
n
though
the
t
r
uth
is
ne
ga
ti
ve
.
It
is
a
ls
o
r
e
f
e
r
r
e
d
to
as
a
type
1
e
r
r
or
.
A
f
a
ls
e
ne
ga
ti
ve
(
F
N
)
is
a
s
it
ua
ti
on
in
whic
h
the
tr
uth
is
pos
it
ive,
but
the
pr
e
diction
is
ne
ga
ti
ve
.
It
i
s
a
ls
o
r
e
f
e
r
r
e
d
to
as
a
type
2
e
r
r
o
r
.
S
ys
tem
pe
r
f
or
manc
e
is
e
va
luate
d
by
c
a
lcula
ti
ng
a
c
c
ur
a
c
y,
p
r
e
c
is
ion,
s
e
ns
it
ivi
ty
(
or
r
e
c
a
ll
)
,
s
pe
c
if
icity,
a
nd
F
-
s
c
or
e
.
In
(
7
)
can
be
us
e
d
to
c
a
lcu
late
the
pr
e
c
is
ion
of
the
s
ys
tem's
pr
e
diction
f
indi
ng
s
:
=
+
+
+
+
(
7
)
In
(
8
)
is
us
e
d
to
de
ter
mi
ne
the
s
ys
tem's
pe
r
f
or
manc
e
,
whic
h
is
e
xpr
e
s
s
e
d
as
the
r
e
c
a
ll
va
lue
or
s
e
ns
it
iv
it
y.
=
+
(
8
)
De
ter
mi
ne
the
s
ys
tem's
pr
e
c
is
ion
in
f
o
r
e
c
a
s
ti
ng
the
s
pe
c
if
ic
type
of
mac
hine
da
mage
by
u
ti
li
z
ing
(
9
)
.
=
+
(
9
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
Ar
ti
f
I
ntell
I
S
S
N:
2252
-
8938
De
s
igni
ng
an
int
e
ll
igent
s
y
s
tem
for
v
ibr
ati
on
diagn
os
is
of
c
e
ntr
if
ugal
w
ater
…
(
W
iw
it
Supr
ihat
ini
ngs
i
h
)
4395
T
he
s
pe
c
if
icity
pa
r
a
mete
r
s
a
nd
F
-
s
c
or
e
a
r
e
c
omput
e
d
us
ing
(
10
)
a
nd
(
11
)
,
r
e
s
pe
c
ti
ve
ly.
=
+
(
10
)
F
-
=
2
2
+
+
(
11
)
3.
RE
S
UL
T
S
AND
DI
S
CU
S
S
I
ON
T
his
int
e
ll
igent
s
ys
tem
s
pe
c
if
ica
ll
y
int
e
nds
to
de
t
e
c
t
vibr
a
ti
ons
in
c
e
ntr
if
uga
l
pumps
by
a
na
lyzing
s
pe
c
tr
um
da
ta.
T
his
can
potentially
mi
nim
ize
the
r
e
quir
e
ment
f
o
r
p
r
of
e
s
s
ional
indi
viduals
to
be
inv
olved
in
the
diagnos
is
pr
oc
e
s
s
.
A
s
ophis
ti
c
a
ted
tec
hnique
w
a
s
de
ve
loped
to
de
tec
t
c
e
ntr
i
f
uga
l
vib
r
a
ti
ons
by
e
x
a
mi
ning
f
r
e
que
nc
y
li
ne
s
in
the
s
pe
c
tr
um
us
ing
a
pr
e
-
e
s
tabli
s
he
d
knowle
dge
ba
s
e
.
Us
e
r
s
can
e
f
f
icie
ntl
y
a
nd
e
f
f
or
tl
e
s
s
ly
a
c
c
e
s
s
s
ophis
ti
c
a
ted
s
y
s
tems
to
a
s
c
e
r
tain
c
e
ntr
if
uge
c
ondit
ions
pr
ompt
ly
.
3.
1.
T
e
s
t
d
at
a
c
oll
e
c
t
ion
of
m
e
as
u
r
e
m
e
n
t
r
e
s
u
lt
s
T
h
e
in
te
l
li
ge
nt
s
ys
te
m
un
de
r
w
e
n
t
tes
t
in
g
ut
i
li
z
i
ng
t
r
a
in
i
ng
da
ta
,
inc
l
ud
in
g
s
pe
c
t
r
u
m
da
ta
,
DE
-
ND
E
p
ha
s
e
be
a
r
i
ng
d
i
f
f
e
r
e
nc
e
s
,
a
n
d
ve
r
ti
c
a
l
ho
r
iz
on
ta
l
p
ha
s
e
d
if
f
e
r
e
nc
e
s
.
T
he
tes
t
d
a
t
a
c
o
mp
r
is
e
s
20
s
e
ts
,
e
a
c
h
c
on
ta
in
in
g
m
is
a
l
i
gn
me
nt
da
ma
ge
da
ta
,
s
ta
t
ic
un
ba
la
nc
e
d
da
mag
e
da
ta
,
dy
na
mi
c
u
nb
a
l
a
nc
e
d
da
mag
e
da
ta
,
l
o
os
e
n
e
s
s
da
ma
ge
da
t
a
,
c
a
vi
ta
ti
on
da
ma
ge
da
ta
,
a
nd
be
a
r
in
g
d
a
ma
ge
da
ta
.
A
m
a
x
i
mu
m
of
1
20
s
a
mp
les
c
a
n
be
g
a
t
he
r
e
d
to
e
v
a
l
ua
te
the
e
f
f
ica
c
y
of
th
e
de
ve
lo
pe
d
in
te
l
li
ge
nt
s
ys
t
e
m
a
pp
li
c
a
ti
on
.
T
e
s
t
d
a
t
a
ge
ne
r
a
t
io
n
i
n
vo
lv
e
s
u
t
il
iz
in
g
p
um
p
f
r
e
q
ue
nc
y
va
l
ue
s
,
b
e
a
r
i
ng
f
r
e
qu
e
n
c
y
va
lu
e
s
,
B
F
P
f
r
e
qu
e
nc
y
v
a
l
ue
s
,
vi
b
r
a
ti
on
di
a
g
nos
i
s
c
h
a
r
ts
,
a
nd
I
S
O
10
81
6
-
13
[
25
]
.
In
a
d
di
t
io
n
,
t
e
s
t
da
ta
w
e
r
e
ga
th
e
r
e
d
f
r
om
th
e
P
91
1B
c
oo
li
ng
wa
te
r
p
um
p
.
H
e
r
e
a
r
e
s
o
me
e
xa
mp
les
o
f
s
pe
c
t
r
um
d
a
t
a
us
e
d
f
o
r
t
e
s
t
i
ng
i
nt
e
l
li
g
e
n
t
s
ys
te
ms
.
F
i
gu
r
e
3
s
ho
ws
th
e
f
a
u
lt
i
nd
ic
a
t
io
n
o
f
F
F
T
r
e
s
ul
ts
,
F
i
gu
r
e
3
(
a
)
de
p
icts
a
s
pe
c
t
r
u
m
da
ta
gr
a
ph
i
l
lus
t
r
a
t
i
ng
th
e
e
f
f
e
c
ts
of
im
ba
lan
c
e
da
ma
ge
.
T
h
e
gr
a
p
h
i
n
di
c
a
t
e
s
a
s
ig
ni
f
ica
nt
m
a
g
ni
tu
de
at
a
f
r
e
q
ue
nc
y
of
2
4
.
17
Hz
,
c
o
r
r
e
s
po
nd
i
ng
to
o
ne
r
e
v
ol
u
ti
on
pe
r
m
inu
te
of
t
he
c
e
n
t
r
i
f
uga
l
p
u
mp
.
A
c
c
or
d
in
g
to
s
o
u
r
c
e
s
,
s
in
gl
e
h
ig
h
a
mp
l
it
ud
e
is
dis
t
in
gu
is
h
in
g
f
e
a
tu
r
e
of
u
nb
a
l
a
nc
e
d
da
ma
ge
.
F
igur
e
3(
b
)
p
r
e
s
e
nts
a
s
pe
c
tr
a
l
da
ta
g
r
a
ph
that
de
mons
tr
a
tes
the
oc
c
ur
r
e
nc
e
of
c
a
vit
a
ti
on
da
mage
.
T
he
gr
a
ph
dis
plays
a
c
ons
picuous
pe
a
k
with
a
l
a
r
ge
magnitude
a
t
a
f
r
e
que
nc
y
of
145
Hz
,
s
ix
ti
mes
the
r
otational
s
pe
e
d
or
f
r
e
que
nc
y
a
t
whic
h
the
blade
s
of
a
c
e
ntr
if
uga
l
pump
pa
s
s
by.
T
he
gr
a
ph
il
lus
tr
a
tes
that
c
a
vit
a
ti
on
da
mage
in
c
e
ntr
i
f
uga
l
pumps
is
c
ha
r
a
c
t
e
r
ize
d
by
a
s
igni
f
ica
nt
a
mpl
it
ude
a
t
6
ti
mes
the
r
e
volut
ions
pe
r
mi
nute
(
R
P
M
)
.
F
igur
e
3(
c
)
pr
e
s
e
nts
a
s
pe
c
tr
u
m
gr
a
ph
il
lus
tr
a
ti
ng
the
e
xtent
of
da
mage
to
the
lo
os
e
ne
s
s
of
the
be
a
r
ing
hous
ing
or
f
ounda
ti
on
.
T
he
gr
a
ph
e
xhi
bit
s
notable
pe
a
ks
a
t
12
.
08,
24
.
17,
48.
33
,
a
nd
72.
5
Hz
.
T
he
f
r
e
que
nc
ies
a
r
e
0.
5,
1,
2,
a
nd
3
ti
mes
the
R
P
M
.
An
e
leva
ted
pe
a
k
in
thi
s
f
r
e
que
nc
y
s
igni
f
ies
a
n
is
s
ue
with
the
f
ounda
ti
on
or
be
a
r
ing.
T
he
pr
e
s
e
nc
e
of
s
igni
f
ica
nt
pe
a
ks
a
t
1X,
2X,
a
nd
3X
va
lues
s
ugg
e
s
ts
th
e
e
xis
tenc
e
of
mi
s
a
li
gnm
e
nt
is
s
ue
s
[
23]
,
[
26
]
,
[
27
]
.
F
igu
r
e
3
(
d)
pr
e
s
e
nts
a
s
pe
c
tr
ogr
a
m
il
lus
tr
a
ti
ng
the
pr
e
s
e
nc
e
o
f
be
a
r
ing
da
mage
.
T
he
g
r
a
ph
e
xhibi
ts
notable
pe
a
ks
a
t
9.
6
7,
58
,
72.
5
,
a
nd
115.
97
Hz
f
r
e
que
nc
ies
.
T
he
f
r
e
que
nc
ies
mentioned
a
r
e
F
T
F
,
B
S
F
,
B
P
F
O
,
a
nd
B
P
F
I
.
I
f
t
he
magnitude
of
a
ny
o
f
the
F
T
F
,
B
S
F
,
B
P
F
O,
or
B
P
F
I
incr
e
a
s
e
s
s
igni
f
ica
ntl
y,
it
s
ugge
s
ts
a
pr
oblem
with
t
he
be
a
r
ing
[
23
]
.
(
a
)
(
b)
(
c
)
(
d)
F
igur
e
3
.
F
a
ult
indi
c
a
ti
on
of
F
F
T
r
e
s
ult
s
of
(
a
)
unb
a
lanc
e
,
(
b)
c
a
vit
a
ti
on,
(
c
)
loos
e
ne
s
s
,
a
nd
(
d)
be
a
r
in
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14,
No.
5:
Oc
tober
2025
:
43
90
-
4402
4396
3.
2.
B
aye
s
ian
n
e
t
wor
k
m
od
e
li
n
g
T
he
de
s
ign
ou
tcome
s
of
the
B
N
w
il
l
s
e
r
ve
a
s
a
be
nc
hma
r
k
f
o
r
de
ve
lo
pin
g
i
ntel
li
ge
nt
s
ys
tem
a
ppli
c
a
ti
o
ns
,
a
s
de
picte
d
in
F
igu
r
e
4
.
T
he
B
N
is
s
t
r
uc
t
ur
e
d
on
t
hr
e
e
hie
r
a
r
c
hica
l
laye
r
s
.
T
he
ini
ti
a
l
s
tage
c
omp
r
is
e
s
a
s
e
t
o
f
c
ha
r
a
c
te
r
is
t
ics
tha
t
r
e
s
u
lt
i
n
de
t
r
i
men
tal
e
f
f
e
c
ts
on
the
p
ump
.
T
he
f
ol
lowi
ng
is
a
c
lar
if
ica
ti
o
n
pr
ov
ided
in
T
a
ble
4.
T
he
s
e
c
ond
-
le
ve
l
a
r
r
a
nge
me
nt
c
om
pr
is
e
s
the
f
a
c
to
r
s
tha
t
r
e
s
ult
in
ha
r
m
,
wh
ic
h
a
r
e
t
he
s
a
me
pa
r
a
mete
r
s
th
a
t
p
r
od
uc
e
da
mage
a
t
lev
e
l
1.
T
a
b
le
5
de
pic
ts
the
or
ga
niza
ti
o
n
a
t
the
s
e
c
o
nd
leve
l
.
Addit
iona
ll
y
,
the
s
e
c
on
d
l
e
ve
l
is
de
te
r
m
ined
us
in
g
vi
br
a
ti
o
n
d
iagn
os
ti
c
s
,
w
hich
a
na
lyze
s
the
c
ha
r
a
c
ter
is
ti
c
s
r
e
s
pons
i
ble
f
or
t
he
da
mage
ident
if
ied
in
b
oth
t
he
f
i
r
s
t
a
nd
s
e
c
o
nd
leve
ls
,
a
s
out
li
ne
d
in
T
a
ble
6
.
B
e
d
a
P
h
a
seH
V
B
e
d
a
P
h
a
se
D
e
N
d
e
1X
2X
3X
0
.
5
X
B
P
F
I
F
T
F
B
P
F
O
B
S
F
B
P
F
BF
H
a
r
m
o
S
t
a
t
i
c
U
n
b
a
l
a
n
c
e
D
y
n
a
m
i
c
U
n
b
a
l
a
n
c
e
M
i
sAl
i
g
n
m
en
t
L
o
o
sen
e
ss
B
e
a
r
i
n
g
C
a
v
i
ta
ti
o
n
F
igur
e
4
.
BN
a
r
c
hit
e
c
tur
e
T
a
ble
4
.
P
r
io
r
p
r
oba
bil
it
y
N
ode
D
e
s
c
r
ip
ti
on
1X
W
he
n
th
e
a
mpl
it
ude
is
la
r
ge
,
a
nd
th
e
f
r
e
que
n
c
y
is
1X
2X
W
he
n
th
e
a
mpl
it
ude
is
la
r
ge
,
a
nd
th
e
f
r
e
que
n
c
y
is
2X
3X
W
he
n
th
e
a
mpl
it
ude
is
la
r
ge
,
a
nd
th
e
f
r
e
que
n
c
y
is
3X
0.5X
W
he
n
th
e
a
mpl
it
ude
is
la
r
ge
,
a
nd
th
e
f
r
e
que
n
c
y
is
0.5X
B
e
da
P
has
e
HV
D
if
f
e
r
e
nt
pha
s
e
be
a
r
in
gs
DE
a
nd
NDE
B
e
da
P
has
e
D
e
N
de
D
if
f
e
r
e
nt
hor
iz
ont
a
l
a
nd
ve
r
ti
c
a
l
pha
s
e
s
B
P
F
I
O
c
c
ur
s
w
he
n
th
e
ma
gni
tu
de
is
e
le
va
te
d
at
th
e
B
P
F
I
f
r
e
que
nc
y
B
P
F
O
O
c
c
ur
s
w
he
n
th
e
ma
gni
tu
de
is
e
le
va
te
d
at
th
e
B
P
F
O
f
r
e
que
nc
y
B
S
F
O
c
c
ur
s
w
he
n
th
e
ma
gni
tu
de
is
e
le
va
te
d
at
th
e
B
S
F
f
r
e
que
nc
y
F
T
F
O
c
c
ur
s
w
he
n
th
e
ma
gni
tu
de
is
e
le
va
te
d
at
th
e
F
T
F
f
r
e
que
nc
y
B
P
F
O
c
c
ur
s
w
he
n
th
e
ma
gni
tu
de
is
e
le
va
te
d
at
th
e
B
P
F
f
r
e
que
nc
y
T
a
ble
5
.
C
ondit
ional
pr
oba
bil
it
y
1
N
ode
D
e
s
c
r
ip
ti
on
H
a
r
mo
H
a
r
moni
c
s
w
it
h
pa
r
a
me
te
r
s
1X
,
2X
,
3X
BF
B
e
a
r
in
g
F
r
e
que
nc
y
w
it
h
s
ympt
oms
of
B
P
F
I
,
B
P
F
O
,
B
S
F
,
F
T
F
T
a
ble
1
.
C
ondit
ional
pr
oba
bil
it
y
2
N
ode
P
a
r
a
me
te
r
S
ta
ti
c
unba
la
nc
e
B
e
da
P
has
e
HV,
B
e
da
P
has
e
D
e
N
de
,
1X
,
C
ondi
ti
ona
l
D
yna
mi
c
unba
la
nc
e
B
e
da
P
has
e
HV,
B
e
da
P
has
e
D
e
N
de
,
1X
,
C
ondi
ti
ona
l
M
is
a
li
ngme
nt
1X
,
2X
,
3X
,
C
ondi
ti
ona
l
L
oos
e
ne
s
s
0.5X
,
1X
,
H
a
r
mo
B
e
a
r
in
g
H
a
r
mo,
BF
C
a
vi
ta
ti
on
H
a
r
mo,
B
P
F
3.
3
.
Re
c
ogn
it
ion
of
in
p
u
t
d
at
a
T
he
int
e
ll
igent
s
ys
tem
identif
ies
input
da
ta
va
lues
by
e
s
tablis
hing
their
va
lue
bounda
r
ies
,
including
the
ove
r
a
ll
vibr
a
ti
on
va
lue
in
R
M
S
,
f
r
e
que
nc
y
s
pe
c
tr
um
da
ta,
DE
ND
E
pha
s
e
dif
f
e
r
e
nc
e
va
l
ue
s
,
a
nd
hor
izonta
l
a
nd
ve
r
ti
c
a
l
pha
s
e
dif
f
e
r
e
nc
e
va
lues
.
T
he
R
M
S
va
lue
a
s
s
e
s
s
e
s
the
pump's
c
ondit
ion
a
c
c
or
ding
to
I
S
O
10816
-
3.
T
a
ble
7
p
r
ovides
a
c
ompr
e
he
ns
ive
r
e
c
or
d
of
the
vibr
a
ti
on
leve
ls
obs
e
r
ve
d
in
c
e
ntr
if
uga
l
pumps
.
T
he
f
r
e
que
nc
y
s
pe
c
tr
um
da
ta
identi
f
ica
ti
on
p
r
oc
e
s
s
is
e
mpl
oye
d
to
a
na
lyze
a
mpl
it
ude
va
lues
a
n
d
e
xtr
a
c
t
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
Ar
ti
f
I
ntell
I
S
S
N:
2252
-
8938
De
s
igni
ng
an
int
e
ll
igent
s
y
s
tem
for
v
ibr
ati
on
diagn
os
is
of
c
e
ntr
if
ugal
w
ater
…
(
W
iw
it
Supr
ihat
ini
ngs
i
h
)
4397
f
r
e
que
nc
y
da
ta
va
lues
,
s
uc
h
as
1X,
2X
,
3X
,
0
.
5X,
B
P
F
,
a
nd
be
a
r
ing
f
r
e
que
nc
ies
.
T
he
da
mag
e
to
the
c
e
ntr
if
uga
l
pump
can
be
de
ter
mi
ne
d
by
c
a
lcula
ti
n
g
the
pr
ior
p
r
oba
bil
it
y
a
nd
c
ondit
ional
p
r
oba
bil
it
y
ba
s
e
d
on
the
identif
ied
f
r
e
que
nc
y.
T
a
ble
8
c
ontains
the
s
pe
c
if
ica
ti
ons
f
or
identif
ying
vib
r
a
ti
on
s
pe
c
tr
um
da
ta.
T
he
mi
nim
a
l
va
lue
indi
c
a
tes
that
the
s
ys
tem
will
not
g
e
ne
r
a
te
e
videnc
e
f
or
BN
unles
s
the
a
mpl
it
ude
s
ur
p
a
s
s
e
s
the
s
pe
c
if
ied
mi
nim
um
th
r
e
s
hold,
as
outl
ined
in
T
a
ble
8.
T
a
ble
7
.
Ove
r
a
ll
vib
r
a
ti
on
a
s
s
e
s
s
ment
O
ve
r
a
ll
V
ib
r
a
ti
on
V
a
lu
e
in
R
M
S
M
a
c
hi
ne
c
ondi
ti
on
0
mm/s
to
<
=
1.4
mm/s
G
ood
c
ondi
ti
on
>
1.4
mm/s
to
<
=
2.8
mm/s
L
ong
-
te
r
m
ope
r
a
ti
on
is
s
ti
ll
pe
r
mi
tt
e
d
>
2.8
mm/s
to
<
=
4.5
mm/s
S
hor
t
-
te
r
m
ope
r
a
ti
on
is
s
ti
ll
pe
r
mi
t
te
d
>
4.5
mm/s
S
to
p
ope
r
a
ti
on;
vi
br
a
ti
on
w
il
l
c
a
us
e
da
ma
ge
T
a
ble
8
.
Vibr
a
ti
on
s
pe
c
tr
um
da
ta
a
s
s
e
s
s
ment
F
r
e
que
nc
y
(
H
z
)
M
in
im
um
a
mpl
it
ude
(
mm
/s
)
1X
(
24.17)
2
2X
(
48.33)
1
3X
(
72.50)
1
0.5X
(
12.08)
0.5
B
P
F
I
(
115.97)
0.3
B
P
F
O
(
77.31)
0.3
B
S
F
(
58)
0.3
F
T
F
(
9.66)
0.3
B
P
F
(
9.66)
0.5
3.
4
.
Dif
f
e
r
e
n
t
p
h
as
e
DE
-
NDE
an
d
H
-
V
a
s
s
e
s
s
m
e
n
t
T
he
identi
f
ica
ti
on
of
DE
NDE
pha
s
e
dif
f
e
r
e
nc
e
va
lues
a
nd
hor
izonta
l
a
nd
ve
r
ti
c
a
l
pha
s
e
d
if
f
e
r
e
nc
e
s
is
us
e
d
to
a
s
c
e
r
tain
s
tatic
a
nd
dyn
a
mi
c
unba
lanc
e
.
If
the
pha
s
e
dif
f
e
r
e
nc
e
be
twe
e
n
the
DE
a
nd
NDE
is
withi
n
the
r
a
nge
of
0º
±
30º
,
it
indi
c
a
tes
the
pr
e
s
e
nc
e
of
s
tatic
unba
lanc
e
.
If
the
d
is
pa
r
it
y
be
twe
e
n
the
DE
a
nd
NDE
is
gr
e
a
ter
than
30º
,
it
s
ugge
s
ts
the
pr
e
s
e
nc
e
of
dyna
mi
c
unba
lanc
e
.
A
ph
a
s
e
dif
f
e
r
e
nc
e
of
90
º
±
30º
be
t
we
e
n
the
ve
r
ti
c
a
l
a
nd
hor
izonta
l
s
ides
of
the
be
a
r
ing
indi
c
a
tes
a
s
igni
f
ica
nt
li
ke
li
hood
of
im
ba
lanc
e
.
A
pha
s
e
dif
f
e
r
e
nc
e
be
twe
e
n
the
ve
r
ti
c
a
l
a
nd
hor
izonta
l
s
ides
of
the
be
a
r
ing
that
f
a
ll
s
outs
ide
the
r
a
nge
of
90º
±
30º
s
ugge
s
ts
a
low
pr
oba
bil
it
y
of
im
ba
lanc
e
da
mage
.
3.
5
.
B
aye
s
ian
n
e
t
wor
k
p
r
ob
ab
il
i
t
y
T
he
M
icr
os
of
t
be
li
e
f
ne
twor
k
ge
ne
r
a
tes
p
r
oba
bil
it
ies
,
s
e
r
ving
as
a
f
r
a
mew
or
k
f
or
de
ve
lopi
ng
int
e
ll
igent
s
ys
tem
a
ppli
c
a
ti
ons
with
M
AT
L
AB
.
P
r
oba
bil
it
y
is
de
r
ived
f
r
om
the
c
ondit
ional
pr
oba
bil
i
ty
table
(
C
P
T
)
.
T
he
C
P
T
dis
plays
the
li
ke
li
hood
of
a
s
pe
c
if
ic
c
ondit
ion
ha
ppe
ning,
given
that
c
e
r
tain
c
ondit
ions
a
r
e
s
a
ti
s
f
ied.
T
he
pr
oba
bil
it
y
of
diagnos
ing
vibr
a
ti
o
n
in
c
e
ntr
i
f
uga
l
pumps
can
be
s
hown
in
T
a
ble
s
9
to
1
3
.
T
a
ble
9
dis
plays
the
C
P
T
of
s
tatic
unba
lanc
e
d
da
mage
.
T
he
P
9114B
c
ooli
ng
wa
ter
pump
da
mage
c
a
us
e
d
by
s
tatic
unba
lanc
e
is
a
f
f
e
c
ted
by
the
pha
s
e
dis
c
r
e
p
a
nc
ies
be
twe
e
n
the
DE
a
nd
ND
E
,
both
hor
izont
a
ll
y
a
nd
ve
r
ti
c
a
ll
y,
a
nd
by
high
a
mpl
it
ude
at
a
pa
r
ti
c
ul
a
r
f
r
e
que
nc
y.
T
he
g
r
e
a
tes
t
li
ke
li
hood
oc
c
ur
s
w
he
n
the
dif
f
e
r
e
nc
e
in
pha
s
e
be
twe
e
n
DE
a
nd
NDE
a
nd
the
dif
f
e
r
e
nc
e
in
pha
s
e
be
twe
e
n
H
a
nd
V
a
r
e
c
ons
t
a
nt,
a
nd
ther
e
is
a
s
igni
f
ica
nt
a
mpl
it
ude
at
the
f
r
e
que
nc
y
of
1X,
pr
ovided
that
the
a
mpl
it
ude
at
1X
is
lar
ge
r
tha
n
at
2X.
T
a
ble
11
pr
e
s
e
nts
the
C
P
T
r
e
s
ult
s
f
o
r
loos
e
ne
s
s
da
mage
.
T
he
da
mage
c
a
us
e
d
by
loos
e
ne
s
s
in
c
e
ntr
if
uga
l
pumps
is
a
f
f
e
c
ted
by
ha
r
moni
c
wa
ve
s
,
pa
r
ti
c
ular
ly
thos
e
with
high
a
mpl
it
ude
at
0.
5
a
nd
1
time
the
main
f
r
e
que
nc
y.
T
he
g
r
e
a
tes
t
li
ke
li
hood
oc
c
ur
s
whe
n
ha
r
moni
c
wa
ve
s
(
Ha
r
mo)
ha
ve
a
high
a
m
pli
tude
at
f
r
e
que
nc
ies
of
0.
5X
a
nd
1X
.
T
he
C
P
T
f
or
mi
s
a
l
ignm
e
nt
da
mage
is
dis
playe
d
in
T
a
ble
11
.
M
is
a
li
gnment
da
mage
in
c
e
ntr
if
uga
l
pumps
is
dis
ti
nguis
he
d
by
s
igni
f
ica
nt
a
mpl
it
ude
s
oc
c
ur
r
ing
at
the
f
un
da
menta
l
f
r
e
que
nc
y
a
nd
its
ha
r
moni
c
s
,
including
twice
a
nd
thr
e
e
ti
mes
the
f
r
e
que
nc
y.
Additi
ona
ll
y
,
t
he
r
e
a
r
e
c
ondit
ional
f
a
c
tor
s
that
mi
ght
in
f
luenc
e
the
e
xtent
of
the
da
mage
.
T
he
maximum
l
ikelihood
oc
c
ur
s
whe
n
the
a
mpl
it
ude
is
e
leva
ted
at
f
r
e
que
nc
i
es
of
1X
,
2X
,
a
nd
3X,
with
the
a
mpl
it
ude
at
1X
be
ing
s
maller
th
a
n
that
at
2X.
T
a
ble
12
dis
plays
the
C
P
T
(
c
a
vit
a
ti
on
da
mage
)
va
lues
.
T
he
e
xtent
of
c
a
vit
a
ti
on
da
mage
to
ma
c
hines
is
de
ter
mi
ne
d
by
the
high
a
mpl
it
ude
of
the
pump's
B
P
F
a
nd
ha
r
moni
c
wa
ve
s
.
T
he
maximum
li
ke
li
ho
od
oc
c
ur
s
whe
n
the
a
mpl
it
ude
is
e
leva
ted
at
the
ba
ndpa
s
s
f
i
lt
e
r
's
f
r
e
que
nc
y,
a
nd
s
e
ve
r
a
l
ha
r
moni
c
wa
ve
s
a
r
e
pr
e
s
e
nt.
T
a
ble
13
dis
plays
the
C
P
T
f
o
r
be
a
r
ing
f
a
il
ur
e
.
T
he
oc
c
ur
r
e
nc
e
of
be
a
r
ing
da
mage
in
c
e
nt
r
if
uga
l
pumps
is
a
f
f
e
c
ted
by
high
a
mpl
it
ude
s
of
s
pe
c
if
ic
be
a
r
ing
f
r
e
que
nc
ies
,
pa
r
ti
c
ular
ly
B
P
F
I
,
B
P
F
O
,
B
S
F
,
F
T
F
,
a
nd
ha
r
moni
c
wa
ve
s
.
T
he
maximu
m
l
ikelihood
oc
c
ur
s
whe
n
the
a
mpl
it
ude
is
e
leva
ted
at
the
f
r
e
que
nc
y
a
s
s
oc
iate
d
with
the
be
a
r
ing
a
nd
other
ha
r
moni
c
wa
ve
s
a
r
e
p
r
e
s
e
nt.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14,
No.
5:
Oc
tober
2025
:
43
90
-
4402
4398
T
a
ble
9
.
P
r
oba
bil
it
y
of
dyna
mi
c
unba
lanc
e
P
a
r
e
nt
node
(
s
)
D
yna
mi
c
unba
la
nc
e
B
e
da
P
ha
s
e
D
e
N
de
B
e
da
P
ha
s
e
H
V
1X
Y
e
s
No
B
a
r
c
ha
r
ts
S
ta
ti
c
S
ta
ti
c
Y
e
s
0.3
0.7
|||||||||||||||
No
0.1
0.9
|||||
D
yna
mi
c
Y
e
s
0.7
0.3
|||||||||||||||||||||||||||||||||||
No
0.2
0.8
||||||||||
D
yna
mi
c
S
ta
ti
c
Y
e
s
0.7
0.3
|||||||||||||||||||||||||||||||||||
No
0.2
0.8
||||||||||
D
yna
mi
c
Y
e
s
0.9
0.1
|||||||||||||||||||||||||||||||||||||||||||||
No
0.3
0.7
|||||||||||||||
T
a
ble
10
.
P
r
oba
bil
it
y
of
l
oos
e
ne
s
s
P
a
r
e
nt
node
(
s
)
M
is
a
li
gnme
nt
1X
2X
3X
Y
e
s
No
B
a
r
c
ha
r
ts
Y
e
s
Y
e
s
Y
e
s
0.95
0.05
|||||||||||||||||||||||||||||||||||||||||||||||
No
0.85
0.15
||||||||||||||||||||||||||||||||||||||||||
No
Y
e
s
0.7
0.3
|||||||||||||||||||||||||||||||||||
No
0.1
0.9
|||||
No
Y
e
s
Y
e
s
0.3
0.7
|||||||||||||||
No
0.1
0.9
|||||
No
Y
e
s
0.1
0.9
|||||
No
0.0
1.0
T
a
ble
11
.
P
r
oba
bil
it
y
of
m
is
a
li
gnment
P
a
r
e
nt
node
(
s
)
M
is
a
li
gnme
nt
1X
2X
3X
Y
e
s
No
B
a
r
c
ha
r
ts
Y
e
s
Y
e
s
Y
e
s
0.95
0.05
|||||||||||||||||||||||||||||||||||||||||||||||
No
0.85
0.15
||||||||||||||||||||||||||||||||||||||||||
No
Y
e
s
0.7
0.3
|||||||||||||||||||||||||||||||||||
No
0.1
0.9
|||||
No
Y
e
s
Y
e
s
0.3
0.7
|||||||||||||||
No
0.1
0.9
|||||
No
Y
e
s
0.1
0.9
|||||
No
0.0
1.0
T
a
ble
12
.
P
r
oba
bil
it
y
of
c
a
vit
a
ti
on
P
a
r
e
nt
node
(
s
)
C
a
vi
ta
ti
on
B
P
F
H
a
r
mo
Y
e
s
N
o.
B
a
r
c
ha
r
ts
Y
e
s
Y
e
s
0.9
0.1
|||||||||||||||||||||||||||||||||||||||||||||
N
o.
0.85
0.15
||||||||||||||||||||||||||||||||||||||||||
N
o.
Y
e
s
0.4
0.6
||||||||||||||||||||
N
o.
0.1
0.9
|||||
T
a
ble
13
.
P
r
oba
bil
it
y
of
b
e
a
r
ing
P
a
r
e
nt
node
(
s
)
B
e
a
r
in
g
H
a
r
mo
BF
Y
e
s
No
B
a
r
c
ha
r
ts
Y
e
s
Y
e
s
1.0
0.0
||||||||||||||||||||||||||||||||||||||||||||||||||
No
0.3
0.7
|||||||||||||||
No
Y
e
s
0.95
0.05
|||||||||||||||||||||||||||||||||||||||||||||||
No
0.0
1.0
3.
6
.
I
n
t
e
ll
igen
t
s
ys
t
e
m
d
e
s
ign
r
e
s
u
lt
s
T
he
c
ons
tr
uc
ti
on
of
an
int
e
ll
igent
s
ys
tem
a
ppli
c
a
ti
on
c
omm
e
nc
e
s
by
int
e
gr
a
ti
ng
a
pr
ogr
a
m
that
a
ll
ows
input
of
da
ta
va
lues
a
nd
da
mage
pr
oba
bi
li
ti
e
s
.
P
r
og
r
a
mm
ing
is
c
onduc
ted
with
in
the
L
iv
e
E
dit
or
s
ubmenu
of
the
M
AT
L
AB
s
of
twa
r
e
.
T
he
a
ppli
c
a
ti
on's
us
e
r
int
e
r
f
a
c
e
include
s
ins
tr
uc
ti
ons
f
or
u
s
ing
the
a
ppli
c
a
ti
on,
a
s
pe
c
tr
um
gr
a
ph
pa
ne
l
f
o
r
dis
playing
the
f
r
e
que
nc
y
s
pe
c
tr
um,
an
input
da
ta
pa
n
e
l,
a
pa
ne
l
f
or
dis
playing
diagnos
is
r
e
s
ult
s
f
or
f
a
il
ur
e
types
that
o
c
c
ur
mor
e
than
50%
of
the
time
ba
s
e
d
on
the
e
nter
e
d
input
da
ta,
a
nd
a
mac
hine
c
ondit
ion
pa
ne
l
that
s
hows
the
mac
hine's
c
ondit
ion
ba
s
e
d
on
the
R
M
S
va
lue.
Up
on
e
ntr
y,
a
gr
a
ph
il
lus
tr
a
ti
ng
the
li
ke
li
hood
of
f
a
il
ur
e
is
pr
e
s
e
nted,
de
picting
the
pe
r
c
e
ntage
pr
oba
bil
it
y
of
da
mage
or
malf
unc
ti
on.
T
he
a
ppli
c
a
ti
on
a
ls
o
c
ontains
mul
ti
ple
c
omm
a
nd
butt
ons
.
T
he
ins
e
r
t
s
pe
c
tr
um
butt
o
n
input
s
vibr
a
ti
on
s
pe
c
tr
um
da
ta.
T
he
s
tar
t
but
ton
ini
ti
a
te
s
the
diagnos
is
p
r
oc
e
dur
e
.
T
he
r
e
s
e
t
butt
on
r
e
s
t
or
e
s
the
int
e
ll
igent
s
ys
tem
a
ppli
c
a
ti
on
to
i
ts
or
igi
na
l
s
tate
,
as
de
picte
d
in
F
igur
e
5.
F
igur
e
5
.
I
ntell
igent
s
ys
tem
a
ppli
c
a
ti
on
dis
play
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
Ar
ti
f
I
ntell
I
S
S
N:
2252
-
8938
De
s
igni
ng
an
int
e
ll
igent
s
y
s
tem
for
v
ibr
ati
on
diagn
os
is
of
c
e
ntr
if
ugal
w
ater
…
(
W
iw
it
Supr
ihat
ini
ngs
i
h
)
4399
3.
7
.
S
ys
t
e
m
t
e
s
t
r
e
s
u
lt
s
Onc
e
the
int
e
ll
igent
s
ys
tem
de
s
ign
is
f
ini
s
he
d,
the
s
ubs
e
que
nt
s
tage
invol
ve
s
tes
ti
ng
the
de
ve
loped
int
e
ll
igent
s
ys
tem
a
ppli
c
a
ti
on.
I
ntelli
ge
nt
s
ys
tem
a
ppli
c
a
ti
on
tes
ti
ng
is
c
onduc
ted
to
ve
r
i
f
y
that
the
s
ys
tem
de
ve
loped
a
li
gns
with
the
r
e
f
e
r
e
nc
e
s
,
a
na
lys
is
,
a
nd
de
s
ign
outcome
s
.
W
it
hin
thi
s
pha
s
e
,
c
ompr
e
he
ns
ive
tes
ti
ng
will
be
c
onduc
ted
on
each
va
r
iant
of
da
mage
d
input
da
ta.
S
ubs
e
que
ntl
y,
the
diagnos
ti
c
outcome
s
ge
ne
r
a
ted
by
the
int
e
ll
igent
s
ys
tem
a
ppli
c
a
ti
on
w
il
l
be
juxt
a
pos
e
d
with
the
manua
l
diagnos
is
,
e
na
bli
ng
the
a
s
s
e
s
s
ment
of
the
e
f
f
ica
c
y
of
the
de
ve
loped
s
ys
tem.
T
we
nty
tes
ts
we
r
e
c
onduc
ted
f
or
each
type
of
da
mage
,
us
ing
dif
f
e
r
e
nt
s
pe
c
tr
um
va
r
iants
,
ove
r
a
ll
v
ibr
a
ti
o
n
va
lues
,
DE
-
NDE
pha
s
e
dif
f
e
r
e
nc
e
va
lues
,
a
nd
h
or
izonta
l
-
ve
r
ti
c
a
l
pha
s
e
va
lues
.
T
he
objec
ti
ve
of
c
on
duc
ti
ng
diver
s
e
tes
ti
ng
is
to
a
s
s
e
s
s
the
e
f
f
icie
nc
y
of
the
in
telli
ge
nt
s
ys
tem
a
ppli
c
a
ti
on
unde
r
dif
f
e
r
e
nt
types
of
da
mag
e
s
c
e
na
r
ios
on
the
P
911B
c
ooli
ng
wa
ter
c
e
ntr
i
f
uga
l
pump.
F
igur
e
6
s
hows
a
gr
a
ph
dis
playing
tes
ti
ng
us
i
ng
input
da
ta
on
dyna
mi
c
unba
lanc
e
d
da
mage
.
T
he
input
da
ta
c
ons
is
ts
of
an
R
M
S
va
lue
of
4.
39
mm
/s
,
a
DE
NDE
pha
s
e
of
83
º
,
a
ve
r
ti
c
a
l
-
hor
izonta
l
pha
s
e
of
140
º
,
a
nd
a
f
r
e
que
nc
y
s
pe
c
tr
um
with
high
a
mpl
i
tude
at
1
r
pm
.
T
he
s
ys
tem
ge
ne
r
a
tes
a
diagnos
is
in
dica
ti
ng
a
dyna
mi
c
unba
lanc
e
d
pr
oblem
with
a
90
%
li
ke
li
hood
a
nd
a
ll
ows
f
or
s
hor
t
-
ter
m
op
e
r
a
ti
ona
l
r
e
c
omm
e
nda
ti
ons
.
T
e
s
ti
ng
wa
s
c
a
r
r
ied
out
to
ve
r
if
y
that
the
in
telli
ge
nt
s
ys
tem
wa
s
ope
r
a
ti
ng
a
c
c
or
di
ng
to
its
de
s
ign
a
nd
to
e
va
luate
the
pe
r
f
or
manc
e
of
the
s
mar
t
s
ys
tem
in
diagnos
ing
the
c
ondit
ion
of
the
c
e
ntr
if
uga
l
pump.
T
e
s
ti
ng
of
the
in
telli
ge
nt
s
ys
tem
p
r
oduc
e
d
120
da
ta.
Af
ter
tes
ti
ng
is
c
ompl
e
te
,
the
da
ta
is
a
na
lyze
d
to
e
va
luate
the
qua
li
ty
of
the
s
mar
t
s
ys
tem
that
ha
s
be
e
n
c
r
e
a
ted.
T
he
tes
t
r
e
s
ult
s
we
r
e
pr
oc
e
s
s
e
d
with
the
he
lp
of
the
c
onf
us
ion
matr
ix
method,
as
s
e
e
n
in
F
igur
e
7.
F
igur
e
6
.
I
ntell
igent
diagnos
is
model
tes
ti
ng
outpu
t
f
or
a
c
ooli
ng‑
wa
ter
c
e
ntr
if
uga
l
pump
F
igur
e
7
.
C
onf
us
ion‑
matr
ix
e
va
luation
o
f
the
p
r
op
os
e
d
f
a
ult
‑
diagnos
is
model
on
the
tes
t
s
e
t
(
n
=
120)
Evaluation Warning : The document was created with Spire.PDF for Python.