TELKOM
NIKA Indonesia
n
Journal of
Electrical En
gineering
Vol.12, No.7, July 201
4, pp
. 5392 ~ 53
9
8
DOI: 10.115
9
1
/telkomni
ka.
v
12i7.514
0
5392
Re
cei
v
ed
No
vem
ber 1
3
, 2013; Re
vi
sed
Jan
uar
y 6, 20
14; Accepted
February 5, 2
014
Resear
ch on Operating Condition Effect on the Shock
Pulse Method
Ruife
ng Yan
g
1
, Jianshe Kang
1
, Lishan Hao
3
, Xinghui Zhang*
1
, Hong
zhi Ten
g
2
,
Haiping Li
1
1
Mechan
ical E
ngi
neer
in
g Col
l
ege, Shi
jiaz
h
u
ang, Ch
in
a
2
Lanzh
ou Eq
ui
pment Mai
n
ten
ance C
enter, L
anzh
ou, Ch
ina
3
Shijiaz
h
u
a
n
g
Arm
y
Comm
an
d Coll
eg
e, Chi
n
a
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: d
y
namic
bnt
@gmai
l
.com
A
b
st
r
a
ct
Roll
in
g b
ear
ing
s
are
o
ne
of th
e
most w
i
d
e
ly
used
el
e
m
e
n
ts in
in
dustri
a
l
ap
plicati
ons.
Sho
ck pu
l
s
e
meth
od
(SPM)
has
prove
n
s
u
ccessfully
as
a
di
agn
ostic to
ol
in
det
ermini
ng
be
arin
g
hea
lth
.
On the
bas
is
of
illustrati
ng
the
pri
n
cip
l
e
of S
P
M, this p
a
p
e
r
mai
n
ly
c
onc
entrates
on
i
n
vestigati
n
g
the
effect of
diffe
rent
oper
ating c
ond
itions o
n
SPM. T
he shock pul
se sign
al
s are
deriv
ed fro
m
the w
i
nd turbin
e gear
box test r
i
g
by SPM
instru
me
nt. T
h
rou
g
h
co
mp
arin
g th
e
slo
p
e
of
dB v
a
lu
es w
h
e
n
th
e rotati
ng
sp
ee
d or
l
oad
ch
an
ges
,
effect of operat
ing co
nditi
ons
on SPM is an
al
y
z
e
d
. T
he an
al
ysis results sh
ow
that SPM is more s
ensitiv
e
to
the rotating s
p
eed i
n
contrast
w
i
th the load.
Ke
y
w
ords
:
sh
ock puls
e
meth
od, cond
itio
n monitor
i
ng,
b
eari
ng fault di
ag
no
sis, operati
ng c
ond
ition
Copy
right
©
2014 In
stitu
t
e o
f
Ad
van
ced
En
g
i
n
eerin
g and
Scien
ce. All
rig
h
t
s reser
ve
d
.
1. Introduc
tion
Wind
en
ergy
is th
e fa
ste
s
t growi
ng
rene
wabl
e
form of en
ergy so
urce i
n
t
he
worl
d.
Ho
wever, hig
h
maintena
nce co
st of wind turbin
e
s
se
riou
sly hamp
e
rs its d
e
velo
pment. As ba
sic
components of
gearbox
es in
drive trains, rollin
g
bearings are widely
us
ed in
wind turbines.
In
indu
strial ap
plicatio
ns, be
aring
s
a
r
e cons
i
dered a
s
critical m
e
chani
cal comp
onent
s and
one
defect may le
ad to cata
stro
phic failu
re of
the wind turb
ines a
nd re
su
lt in major economi
c
losse
s
.
Con
d
ition mo
nitoring i
s
an effective way to
detect the defect
s
and p
r
event the failure
s of
beari
n
g
s
. Ma
ny re
sea
r
che
r
s h
ad
re
sea
r
ched th
e dia
g
n
o
si
s meth
od
of bea
ring
s [1
-4]. A revie
w
of
vibration a
n
d
aco
u
sti
c
me
asu
r
em
ent m
e
thod
s for th
e dete
c
tion in
rolling
bea
ri
ngs i
s
p
r
e
s
en
ted
for bea
ring
s condition mo
ni
toring by Tan
don an
d Ch
o
udhu
ry [5].
Shock p
u
lse
method
(SPM
) ha
s b
een
wi
dely us
ed a
s
a qua
ntitative method fo
r b
earin
gs
con
d
ition m
o
nitoring
[6]. SPM is
de
rived
from SPM
in
strum
ent
whi
c
h
develo
ped
and
pate
n
te
d in
the ea
rly 70’
s in S
w
ed
en
[7]. Throu
g
h
years
of t
e
sting
and
rese
archin
g, SPM has
be
en
develop
ed a
n
d
pe
rfecte
d so as to repre
s
ent the
“true
”
op
eratin
g
condition
of th
e bea
rin
g
bei
ng
monitored [8-12].
SPM can pro
v
ide a direct
sho
c
k value indicating the beari
ng condi
tion [13]. By collectin
g
the sho
ck v
a
lue
s
un
de
r different
op
erating
condi
tions, the
trend
of sh
ock valu
es
ca
n be
analyzed to
predi
ct the
o
peratin
g
con
d
ition of b
ear
ings
bein
g
te
sted. In thi
s
pape
r, we m
a
inly
investigate th
e effect of
ro
tating sp
eed
and
lo
ad
on t
he SPM, whichh
as
not b
e
en research
e
d
.
The
sho
c
k
pu
lse
sig
nal
s a
r
e obtai
ned
by
SPM in
strum
ent from
two
experim
ents
unde
r diffe
re
n
t
operating con
d
itions.
The structu
r
e
of this paper
is org
ani
zed
as fo
llows. In Section 2, the theorie
s of SPM are
illustrated bri
e
fly. The assessm
ent parameters
of SPM
are di
scussed.
In Section 3,
th
e
experim
ents
are introdu
ce
d. In
Section 4, the experimental re
su
lts are an
alyzed. Sectio
n
5
discu
s
ses the
analysi
s
re
su
lts and con
c
lu
sion
s are given in Section
6.
2.
The Thories
of SPM
SPM is th
e
monitori
ng
an
d an
alyzatio
n
of
hig
h
frequ
ency
sh
ock
wave
s g
ene
rated by
a
beari
n
g
whil
e
rotatin
g
. Wh
en
rolling
el
e
m
ent
conta
c
t
s
with
a d
a
m
aged
a
r
ea
of
ra
ce
way i
n
t
h
e
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Re
sea
r
ch on
Operating Co
ndition Effect on the Shock
Pulse Meth
od
(Ruifen
g
Yan
g
)
5393
rotating b
e
a
r
i
ng, high freq
uen
cy sh
ock
wave
s are
ge
nerate
d
in th
e interfa
c
e b
e
t
ween th
e roll
ing
element
and
the racewa
y. The sho
c
k
waves,
which present
non
-continu
ous pul
se
s, are
transfe
rred
to
the
bea
ring
hou
sing
thro
ugh th
e i
nne
r materi
al
of b
earin
g. Th
e
signal
s of
sho
c
k
wave
s are re
ceived
by the sho
c
k pul
se
sen
s
o
r
s
whi
c
h are
pla
c
ed
on the be
ari
n
g hou
sing. T
h
e
way the
s
e
si
gnal
s a
r
e
se
parate
d
i
s
re
ally what
ma
ke
s thi
s
te
ch
nology u
n
iqu
e
. Unli
ke vib
r
ation
analysi
s
that monitors
a broad
vibration band and
t
h
e
n
trie
s to isol
ate uniq
ue freque
nci
e
s, S
P
M
has
develop
e
d
a mea
n
s to
only “loo
k”
a
t
the high fre
quen
cy sig
n
a
l
s of rotatin
g
beari
n
g
s
. Th
e
sho
c
k pul
se
sen
s
o
r
s
are
also diffe
re
nt from
othe
r vibration
sensor
s an
d
operate at their
resona
nce freque
ncy of 3
2
kHz. Acco
rd
ingly the
resulting bea
rin
g
sign
als a
r
e
strong
est. T
h
e
sho
c
k pul
se
s
will cause the
dampe
d o
sci
llation of t
he
sen
s
o
r
s
at their natu
r
al fre
quen
cy [5]. The
amplitude of
these
sign
als is pro
portio
n
a
l to
the dam
age level of the bea
ring
s.
Then the b
a
n
d
-
pass filter ca
n filter the me
cha
n
ical interferen
ce n
ear
the bea
ring a
nd only let the high fre
que
ncy
comp
one
nt pass. The re
su
lting high freq
uen
cy si
gn
als pass thro
ugh
the amplifier
and then afte
r
envelop
e de
modulatio
n [14], the shock
values in
dicat
i
ng the bea
rin
g
con
d
ition can be de
rived
.
The ab
solute
sho
ck p
u
lse
level of a bearin
g, mea
s
ured in
dB
sv
,
is both a function of
rolling
sp
eed
and of b
eari
n
g co
ndition. T
o
neutraliz
e the effect of
rolling
spee
d
on the me
asu
r
ed
value, SPM t
hen
cal
c
ul
ate
the initial
va
lue
dB
i
, th
e
starting
p
o
int of the
condi
tion scal
e fo
r a
particula
r bea
ring. The
dB
i
value dep
end
s on the
size
of inner
ra
ce
and rotating speed, which i
s
defined a
s
[6]:
2
0
0
6
2150
i
d
B
(
l
gN
.
l
gD
l
g
)
(1)
Whe
r
e
N
den
otes the rotating sp
eed of the bea
ring,
D
is the inne
r di
ameter.
The co
ndition
scal
e
is grad
ed in normali
zed
sho
c
k value,
dB
n
, which is defined as:
06
200
0
20
.
SV
dB
n
l
g
ND
(2)
Whe
r
e
SV
de
n
o
tes the sho
c
k value.
SPM can
sa
mple the
sho
ck
pul
se am
p
litude over
a
perio
d of time
and di
splay
s
: (1) Th
e
maximum value
dB
m
for the small num
b
e
r of stron
g
s
hock pul
se
s. (2) Th
e ca
rp
et value
dB
c
for
the large n
u
m
ber of we
aker sho
ck p
u
lses.
The maxim
u
m value
dB
m
define
s
the
beari
ng’
s po
sition o
n
the
con
d
ition
scale. The
differen
c
e bet
wee
n
dB
m
and
dB
c
is
used
for an
alysi
s
of the cau
s
e
s
for redu
ce
d
or b
ad
con
d
ition.
This meth
od i
s
used for
co
mparative rea
d
ing on different beari
n
g
s
.
The m
a
jo
r b
enefit of SP
M is p
r
ovidin
g a
dire
ct i
n
dicatio
n
of
b
earin
g
con
d
ition o
n
a
Gree
n-Yell
ow-Re
d
scale.
Gree
n mea
n
s a good be
ari
ng, Yellow is
a beari
ng wit
h
early dam
a
ge
and
Red i
s
m
o
re
seve
re d
a
mage. T
h
is i
s
very imp
o
rt
ant wh
en m
o
nitoring i
n
stal
led be
arin
gs
on
ma
c
h
in
es
for w
h
ic
h th
er
e a
r
e
no
tren
d
s
or
c
o
mp
a
r
ab
le
r
e
a
d
in
g
s
.
T
h
es
e p
a
r
a
me
ters fo
r
evaluating th
e beari
ng con
d
ition ca
n be
illustrate
d as
Figure 1.
Figure 1. The
Diagram of Param
e
ters
for Evaluating the Bearin
g Co
ndition
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 7, July 201
4: 5392 – 53
98
5394
3.
Wind Turbin
e
Gearb
o
x Experiment
Figure 2 sho
w
s the
win
d
turbin
e test ri
g use
d
for d
a
t
a colle
ction.
The test ri
g i
n
clu
d
e
s
two ge
arb
o
xe
s, one i
s
spe
ed-d
o
wn ge
a
r
box, anot
h
e
r is spee
d-u
p
gearbox, a 1
000
kw m
o
tor is
use
d
for d
r
iving the
ge
arboxes, th
ree
motors
fo
r l
oadin
g
with
power
355
kw, 500
kw,
655
kw
respe
c
tively, and a spee
d
and torque
sensor for m
e
asu
r
ing th
e rotating sp
eed
and torq
ue. The
spe
ed-up ge
a
r
box is the te
st gear
box. In the test gearbox, the ring
gear i
s
statio
nary, a sun g
ear
r
o
ta
te
s
ar
o
und
a
fi
xed ce
nter, and plan
et gears not onl
y rota
te arou
nd their own cente
r
s b
u
t also
revolve aroun
d the ce
nter
of the sun
ge
ar. T
he pl
ane
t gears me
sh
simultan
eou
sly with both
the
sun
ge
ar
and
the
ring
gea
r. The
inn
e
r
stru
ctur
e of t
he te
st ge
arb
o
x and
mea
s
uring
poi
nts
of
SPM sen
s
or
are de
picte
d
as Figu
re 3.
Figure 2. The
Diagram of the Gea
r
box T
e
st Rig of FL
600 Wi
nd Tu
rbine
Figure 3. The
Diagram of the Test
G
earbox and Mea
s
uri
ng Point
s
Location
As sh
own in
Figure 3, the test gearbox
has
fou
r
sh
a
fts, includin
g
low spee
d sh
aft and
high spee
d shaft which are viewed a
s
i
nput and o
u
tput shaft re
spectively, su
n gea
r sh
aft and
intermediate
shaft. Planet gear i
s
driven by
low speed shaft, meshing
with sun gear. Sun
gear
mesh
es
with
gear
of intermediate
shaf
t. Gear of hi
gh sp
eed
sh
aft also me
shes
with ge
a
r
of
intermediate s
haft.
In this
pap
er,
we
a
rra
nge
d two
expe
ri
ments. T
he fi
rst i
s
no l
o
a
d
expe
rime
nt but the
spe
ed i
n
crea
se
s from 5
0
0
rpm to
15
17rp
m
with
an in
creme
n
t of 25
0
rpm
.
The
se
co
n
d
experim
ent is the co
nsta
nt spe
ed at 1
5
17rp
m
an
d five different lo
ad level
s
: 0, 25%, 50%, 7
5
%,
and 10
0%. In each exp
e
ri
ment, six measu
r
ing p
o
int
s
(a
s shown in Figure 3) a
r
e cho
s
e
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Re
sea
r
ch on
Operating Co
ndition Effect on the Shock
Pulse Meth
od
(Ruifen
g
Yan
g
)
5395
4. Experimenta
l
Results
The value
s
of
dBm
and
dBc
at ea
ch me
asu
r
ing
point
unde
r differe
nt workin
g co
ndition
s
are d
e
rived b
y
SPM instru
ment. Figure 4-9
sho
w
the
experim
ental
results at different me
asuri
ng
points.
Figure 4. The
Diagram of
dB
Values at the Mea
s
u
r
ing
Point 1
Figure 5. The
Diagram of
dB
Values at the Mea
s
u
r
ing
Point 2
Figure 6. The
Diagram of
dB
Values at the Mea
s
u
r
ing
Point 3
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 7, July 201
4: 5392 – 53
98
5396
Figure 7. The
Diagram of
dB
Values at the Mea
s
u
r
ing
Point 4
Figure 8. The
Diagram of
dB
Values at the Mea
s
u
r
ing
Point 5
Figure 9. The
Diagram of
dB
Values at the Mea
s
u
r
ing
Point 6
Acco
rdi
ng to
Figure 4
-
9,
the values of
dB
m
and
dB
c
chan
g
e
obviou
s
ly with th
e
increa
sing
sp
eed. Ho
weve
r, whe
n
the
l
oad is i
n
crea
sing, the
cha
nge of
dB
m
and
dB
c
values is
not evide
n
t. In o
r
de
r to
qua
ntitatively explai
n thi
s
p
r
o
b
lem,
norm
a
lization
processing
of
hori
z
ontal o
r
dinate is imp
l
emented a
n
d
the slope
of the
dB
values at ea
ch
measuri
ng p
o
int
unde
r differe
nt spee
d and
load is
cal
c
ul
ated.
Table 1
and Tabl
e 2 p
r
esent the re
sults.
Thro
ugh
co
m
pari
s
on
of th
e slo
pe, an
al
ysis
re
sult sh
owsthat the
rotating spee
d
has
a
greate
r
im
pa
ct on
the
dB
value
s
tha
n
load. Th
erefo
r
e, SPM i
s
more
sen
s
itive to the
rotat
i
ng
spe
ed in cont
rast
with load
.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Re
sea
r
ch on
Operating Co
ndition Effect on the Shock
Pulse Meth
od
(Ruifen
g
Yan
g
)
5397
Table 1. The
Slope of
dB
Value
s
with No
-load a
nd Ch
angin
g
Spee
d
Measuring point
Slope
dB
m
dB
c
1 15
33
2 31.5
42
3 25.5
31.5
4 27
42
5 27
28.5
6 25.5
24
Table 2. The
Slope of
dB
Value
s
with Co
nstant Spee
d
and Ch
angi
n
g
Load
Measuring point
Slope
dB
m
dB
c
1 3
5
2 -6
-2
3 4
5
4 2
5
5 1
9
6 7
0
5. Discus
s
ions
Rotating
spe
ed and l
oad
a
r
e both th
e b
a
si
c data of
o
peratin
g cond
itions in b
e
a
r
ings. But
according
to
analysi
s
re
su
lt, in cont
ra
st with the
loa
d
, SPM are
more
sen
s
itive to the
rotat
i
ng
spe
ed.
As mention
e
d
earli
er, SPM need
s the
rotating
spe
e
d
and in
ner
d
i
ameter a
s
in
put data.
Then i
n
itial v
a
lue
dB
i
, whi
c
h depe
nd
s on
the sp
eed
and
i
nne
r
di
ameter, ca
n be calculated
to
obtain th
e
no
rmali
z
ed
sho
c
k value
dB
n
. The
dB
n
can
indicate
the
b
earin
g con
d
ition
a
n
d
dB
m
and
dB
c
values a
r
e de
rived b
a
se
d on the
dB
n
. Therefore, the SPM can be
co
me a quantit
ative
method for d
e
tecting the d
e
fects of be
arings.
Obviou
sly, the two i
nput
data which
SPM need
s
are i
r
relevant
to the loa
d
, but are
clo
s
ely rel
a
te
d to the rotating spee
d. Th
erefo
r
e, t
he a
nalysi
s
re
sult
that SPM is
more
se
nsitiv
e to
rotating spee
d than load i
s
explainable.
6. Conclu
sion
The u
s
e of shock pul
se te
chni
que
s ha
s proven
rema
rka
b
ly su
cce
s
sful as a
qua
ntitative
method for
b
earin
g conditi
on monito
ring
. The prin
cipl
e of sho
c
k p
u
lse m
e
thod
and a
s
sessm
ent
para
m
eters o
f
SPM are illu
strated. T
h
is
pape
r mainly
con
c
e
n
trate
s
on inve
stigating the effe
ct of
operating
con
d
ition on SP
M throug
h co
mpari
ng the
dB values
wh
en the rotatin
g
spe
ed o
r
lo
ad
cha
nge
s. The
analysi
s
result sho
w
s tha
t
SPM is
more se
nsitive to
the rotating
speed in
co
ntrast
with the load.
Referen
ces
[1]
Z
hang
YP, Z
h
ang
T
,
T
eng J, Su
HS. Impro
v
ed E
n
sem
b
le
Empiric
a
l
Mo
de
Decom
pos
ition
for ro
lli
n
g
bear
ing
fault
di
agn
osis
.T
ELK
O
MNIKA Indon
esia
n Jo
urna
l
of Electrica
l
En
gin
eeri
n
g
. 2
0
1
4
; 12(
1): 53
9-
549.
[2]
Li H, Yin YJ. Beari
ng fault d
i
a
gnos
is bas
ed o
n
Lap
lac
e
W
a
velet T
r
ansform
.T
ELKOMNIKA Indon
esi
a
n
Journ
a
l of Elec
trical Eng
i
ne
eri
n
g
. 201
2; 10(8)
: 2139-2
1
5
0
.
[3]
Lin
g
X, W
e
i C.
Stud
y
of Intel
l
i
ge
nce
Di
agn
o
s
is s
y
stem for
w
i
nd
turb
in
e g
earb
o
x
fa
ult.
TEL
K
OMNIKA
INDONESIAN Journ
a
l of Elec
trical Eng
i
ne
eri
n
g
. 201
3; 11(8)
: 4691-4
6
9
7
.
[4]
F
eng Y
X
. Stud
y of fau
l
t dia
g
n
o
sis metho
d
fo
r
w
i
nd tur
b
in
e
w
i
t
h
dec
isio
n c
l
assificati
on
al
gorithms a
n
d
expert s
y
stem.
T
E
LKOMNIKA Indon
esi
an Jou
r
nal of Electric
al Eng
i
ne
eri
n
g
.
2012; 1
0
(5): 9
05-9
10.
[5]
T
andon N, C
h
oud
hur
y A. A r
e
vie
w
of vibr
ati
on a
nd
acousti
c measur
emen
t methods for t
he d
e
tectio
n
of defects in rol
ling
elem
ent be
arin
gs.
T
r
ibolo
g
y Internatio
na
l
. 1999; 3
2
: 46
9–4
80.
[6
]
L
i
Z, H
e
ZJ, Zi
YY, C
h
en
X
F
. Be
a
r
i
n
g
co
n
d
i
t
i
o
n mo
ni
to
ri
ng
ba
se
d on
sho
ck pu
l
s
e me
th
od
a
nd
improve
d
red
u
nda
nt lifting sc
heme.
Mathe
m
atics and C
o
mputers in Si
mul
a
tion
. 20
08; 79
: 318-33
8.
[7]
T
he shock p
u
lse m
e
thod
for determi
ni
n
g
con
d
itio
n o
f
anti-friction
bear
ings.
SP
M T
e
chnic
a
l
Information
. Sw
eden: SPM Instruments AB.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 7, July 201
4: 5392 – 53
98
5398
[8]
Masaru F
,
Mi
n
o
ru K. SPM s
h
ock pu
lse m
e
thod f
o
r di
ag
no
sitic s
y
stem of
rotating
rol
l
b
eari
ng.
Japan
T
appi Jo
urn
a
l.
200
3; 57(4): 52
-57.
[9]
Moran
do
LE.
Measuri
n
g
sho
ck puls
e
s is
id
eal for
b
eari
ng
cond
ition
mon
i
torin
g
.
Pulp Paper
. 1
988;
6
2
:
96–
98.
[10]
Moran
do L. Me
asuri
ng Sh
ock Pulse.
SPM Instrum
e
nt Inc
. 2005.
[11]
T
andon N,
Ya
dava
GS, Ram
a
krish
na KM.
A comp
ariso
n
of some
co
ndit
i
on
mon
i
torin
g
techni
qu
es f
o
r
the detecti
on o
f
defect in in
du
ction motor b
a
l
l
bear
in
gs.
Me
chan
ical Syste
m
s a
nd Si
gn
al
Processin
g
.
200
7; 21(1): 24
4-25
6.
[12] Prabh
u
R.
Rol
l
i
ng elem
ent dia
gnostics.
Pr
oc
eed
ing
of th
e I
ndo-
US Sy
mp
o
s
iu
m o
n
E
m
erg
i
ng
T
r
ends
i
n
Vibrati
on a
nd
Noise E
ngi
ne
e
r
ing
. Ne
w
Del
h
i
.
1996; 31
1–3
2
0
.
[13]
Butler
DE. T
he shock
pu
lse
method
for
th
e
detecti
on
of d
a
mag
ed r
o
ll
in
g
be
arin
g.
NDT
Int.
197
3; 6:
92–
95.
[14]
Ran
dal
l RB, A
n
toni J. Ro
lli
ng
eleme
n
t be
ari
ng di
ag
nostics
- A tutorial.
Me
chan
ical Syste
m
s a
nd Si
gn
a
l
Processi
ng
. 20
11; 25: 48
5–5
2
0
.
Evaluation Warning : The document was created with Spire.PDF for Python.