TELKOM
NIKA Indonesia
n
Journal of
Electrical En
gineering
Vol. 14, No. 3, June 20
15, pp. 434 ~ 4
4
0
DOI: 10.115
9
1
/telkomni
ka.
v
14i3.796
7
434
Re
cei
v
ed Ma
rch 4, 2
015;
Re
vised
Ma
y 10, 2015; Accepted Ma
y 24
, 2015
Wavelet
Transform Based Fault Diagnosis of BLDC
Motor Drive
Ram
e
s
h
Bal
a
ji S.M.*
1
, M
uniraj C
2
, Mekala N
3
K.S.Rangas
am
y Co
lle
ge of T
e
chno
log
y
, T
i
ru
chen
go
de, Nam
a
kkal, T
a
milna
du, India
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: rameshb
a
la
ji
13@
gmai
l.com
1
, c.muniraj@g
m
ail.com
2
,
natara
j
meka
la
@gmai
l
.com
3
A
b
st
r
a
ct
Due to t
he d
e
vel
o
p
m
e
n
t in
scienc
e a
nd
techno
lo
gy tre
m
e
n
d
ous i
m
pr
ove
m
e
n
t in s
o
lid stat
e
devic
es an
d circuits, so w
e
nee
d to reduc
e the co
mp
l
e
xi
ty of control circuits. T
he Brushless DC M
o
t
o
r
(BLDCM)
has
simple
co
nstru
c
tion, stator c
o
nsists of
w
i
n
d
i
ng
an
d rotor
c
onsists
of p
e
r
m
a
n
e
n
t
ma
gne
t. It
has hi
gh
er efficiency a
nd h
i
g
h
spee
d ran
g
e
.
In this
paper
prese
n
ts the vibratio
n an
alysi
s of single p
h
a
s
e
ope
n circuit fa
ult cond
itions
and the si
gn
al
s are extr
acte
d from the Dis
crete W
a
velet
T
r
ansform (D
W
T
).
T
he Sp
artan-
6
F
P
GA process
o
r w
a
s i
m
pl
emented
d
u
e
to
th
eir
hig
h
perfor
m
a
n
ce
of th
e c
ontrol
a
ppl
icati
o
n
s
in rece
nt deca
des. In this pr
ocess Sin
u
so
i
dal Pu
lse W
i
dt
h Modu
lati
on (
SPW
M) techni
ques w
e
re us
e
d
to
control th
e sp
e
ed of B
L
DC M
o
tor. T
he ex
pe
rimental
an
alys
is is carr
ied
ou
t in 3
phas
e, B
L
DCM
drive. T
h
e
vibrati
on sig
n
a
l
s are me
asur
e
d
by usin
g Acceler
o
meter thro
ugh L
ab VIEW
softw
are.
Ke
y
w
ords
:
brush
l
ess DC motor
(BLD
C
M),
vibratio
n ana
l
ysis, discr
ete wavelet transfor
m
(DWT
), Lab
VIEW
Copy
right
©
2015 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
A motor that retain
s the chara
c
te
risti
c
s of
a dc motor it eliminates the commut
a
tor and
bru
s
he
s i
s
ca
lled a B
r
u
s
hl
ess
DC
Moto
r (BL
D
CM). It
ha
s stato
r
a
nd rotor,the
stator carrie
s t
h
e
armatu
re
an
d the rotor
carri
es th
e pe
rman
ent ma
g
net, it rotor
and a
r
matu
re rem
a
in
s
static.
BLDCM can i
n
many case
s re
pla
c
e
con
v
entional
DC
motors. The f
unctio
n
of the
comm
utator i
s
to coll
ectio
n
of cu
rrent fro
m
the
arm
a
ture
c
ond
ucto
r a
nd it
co
n
v
erts th
e alt
e
rnatin
g
cu
rrent
indu
ced i
n
th
e arm
a
ture condu
ctors int
o
unidi
re
ct
ion
a
l cu
rrent in t
he external l
oad
circuit. A
dc
curre
n
t is passing throu
gh the bru
s
hes be
ca
use
of commutator and b
r
u
s
he
s a
c
tion and
unidirectio
nal
torque i
s
d
e
v
eloped in
armature
co
nd
uctor. T
he B
L
DC moto
r
with wi
nding
s in
delta co
nfigu
r
ation give
s low torq
ue a
t
low rpm b
u
t can give
highe
r ran
g
e
s
of rpm. T
h
e
config
uratio
n
gives
high t
o
rqu
e
at lo
w rpm
but
the
motor
ca
nn
ot be o
perated in
high
er
rpm
rang
es.
BLDCM havi
ng roto
r po
siti
on se
nsor, it conve
r
ts the i
n
formatio
n a
bout the roto
r position
into a suitabl
e elect
r
ical signal
s, hall e
ffect se
n
s
o
r
is used in thi
s
motor fo
r sense the rot
o
r
pos
ition.It having more f
e
atures
lik
e
, low
maintenanc
e bec
a
use of
no brus
hes
,
long life,
low
inertia, high e
fficiency. In [8], propo
sed
the wave
let theory for d
e
velopin
g
a se
n
s
orl
e
ss d
r
ive for
BLDC m
o
tors. Wavelet the
o
ry is broadly
used in
an i
m
age p
r
o
c
e
s
sing that en
a
b
les to an
alyze
non
stationa
ry signal
s in freque
ncy an
d
time domai
n.
Referen
c
e [1
4], discusse
d
the over
cu
rrent
prote
c
tion fo
r Perma
nent
Magnet Brushl
ess
DC Motor (PM
B
LDCM) op
erating u
nde
r the
different o
perating conditio
n
s at
no lo
ad
to full
load
o
peratio
n. The
mea
s
ureme
n
t of the in
stant
freque
ncy is
a signifi
cant feature
of the anticip
ated fa
ult detection
algorith
m
[5].
In [9], proposed
the
Simulati
on
Mo
delin
g
of
BL
DCM
drive and
harm
oni
c
a
n
a
lysis
of current, spee
d and
accele
ration
usin
g Di
scret
e
Wavelet Transfo
rm Te
chniqu
e. DWT
is a useful t
ool to detect
the
disturban
ce
s
of instant time varyi
ng sig
nals. In [10] mech
ani
cal vibr
ation i
s
dire
ctly related to
the
large
torque
ripple. Meth
od
s of
effective
voltage ve
cto
r
combi
nation
s
to
re
du
ce th
e torque
rip
p
l
e
and thereby redu
ce the me
cha
n
ical vibra
t
ion.
In this main
work i
s
to a
nalysi
s
the vibration
sign
als u
s
ing
DWT, accel
e
ro
meter is
placed i
n
BLDCM
drive; it will
measure the vibr
ation signal
s by
using
Lab VIEW. The
NIcDAQ
card is u
s
ed
to measu
r
e the real time
sign
als
like voltage, cu
rre
n
t and vibrati
on. This pa
p
e
r is
con
s
id
ere
d
b
r
ief abo
ut di
screte
wavel
e
t trans
f
o
rm
in se
ction
2, Section 3
discu
s
sed t
he
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Wa
velet Tran
sform
Based
Fault Diag
no
sis
of BL
DC
Motor Drive (Ram
esh Bal
a
ji.S.M)
435
LabVIEW. Section 4 di
scu
s
sed
the block diag
ra
m of BLDCM
drive. Section 5 discussed
Experimental
Re
sult and
se
ction 6 is
con
c
ludi
ng com
m
ents.
2. Discre
t
e
Wav
e
let Transform
In continuo
u
s
wavelet transfo
rm (CWT),
the m
o
ther is dil
a
ted and transl
a
ted
contin
uou
sly over
a real continuo
us nu
mber sy
ste
m
and it can g
enerate sub
s
tantial red
u
n
dant
informatio
n. The di
screte
Fouri
e
r t
r
an
sform
(D
FT
)
is obtai
ned f
r
om the F
ouri
e
r tra
n
sfo
r
m
by
repla
c
in
g
the
integral
with
a
finite sum. The
di
sc
rete wavelet
transform (DWT
) is
obtained from
the wavel
e
t transfo
rm in
the sam
e
way. The
r
e
exists a m
o
re su
btle wa
y to perform
the
decompo
sitio
n
u
s
ing
wave
lets. Th
ese
si
gnal
s A
and
D a
r
e
inte
re
sting but
we g
e
t 200
0 val
u
es
instea
d of the 1000
we h
ad then loo
k
i
ng ca
refu
lly at the comp
u
t
ation we ma
y keep only
one
point o
u
t of two i
n
ea
ch
o
f
the two
200
0-len
g
th
sam
p
les to get th
e complete
i
n
formatio
n T
h
e
pro
c
e
ss
on t
he rig
h
t whi
c
h incl
ude
s d
o
wn
sam
p
lin
g, prod
uces
DWT coeffi
ci
ents. Thi
s
is the
notion of do
wn samplin
g. We produ
ce two se
que
nce
s
called
cA and
cD.T
o gain a bet
ter
appreci
a
tion
of this
pro
c
e
s
s let'
s
perfo
rm a o
n
e
-
st
a
g
e
di
screte
wa
velet tran
sform of a
sig
nal.
In
this
sign
als
will b
e
a
pu
re
sinu
soi
d
with hi
gh-f
r
e
quen
cy n
o
ise ad
ded
to
it. Sample
DWT
coeffici
ent sig
nal develop
e
d
block is
sho
w
n in Figu
re
1.
Figure 1. Sample DWT Coefficient
s Signal
s
In these sig
n
a
ls wa
s ge
ne
rated u
s
ing
MATLAB cod
e
neede
d an
d the signal
s
were su
b
divided into lo
wer frequ
en
cy (cA) an
d higher frequ
en
cy (cD).
s = si
n(20.*lin
spa
c
e
(
0,pi,10
00)) + 0.5.*ra
nd(1,1
000
);
[cA,cD] = d
w
t
(
s,'d
b2'
);
Here, db2 is t
he name of th
e wavelet we want to use for the analy
s
i
s
.
Notice the de
tail coeffici
ent
s cD is
sm
all
and con
s
ist
mainly of a hi
gh-frequ
en
cy noise,
while the a
p
p
r
oximation
co
efficients
cA contai
ns
mu
ch less noi
se than do
es the
origin
al sig
nal
.
[length (cA
)
length(cD)]
3. Lab VIEW
Lab VIEW
(L
aboratory Virt
ual Inst
rume
n
t
E
nginee
ring
Wo
rkben
ch
) prog
ram
s
are
calle
d
virtual in
stru
ments
be
cau
s
e th
eir
app
e
a
ran
c
e
an
d o
peratio
n imita
t
e physi
cal i
n
strum
ents su
ch
as
multimete
r
s an
d o
s
cill
oscop
e
s.
La
b VIEW
co
n
t
ains a com
p
reh
e
n
s
ive set
of
tool
s
f
o
r
displ
a
ying, a
c
quirin
g
, an
alyzing
an
d
stori
ng d
a
ta a
s
well a
s
tool
s to
help
you
tro
uble
s
ho
ot co
de
you write.
Data a
c
q
u
isit
ion
is the p
r
o
c
e
ss
of sam
p
li
ng signal
s that
measure
real worl
d
p
h
ysical
con
d
ition
s
a
nd convertin
g
the re
sulti
ng samp
le
s
into digital n
u
meri
c value
s
that can
be
manipul
ated
by a comp
ute
r
, BLDC m
o
tor
with vibr
ation
sen
s
o
r
u
s
i
ng La
b VIEW is
sho
w
n
in t
he
Figure 2.
NI 923
4 is
a
nalog
DAQ
whi
c
h h
a
s fo
ur chan
nel
s, 24 bit re
sol
u
tion; it can
o
perate i
n
temperature
range f
r
om
-4
0.c to 70.
c ra
nge.
Th
e NI
9234
anal
og i
nput chan
nel
s a
r
e referen
c
ed
to ch
assi
s g
r
ound
thro
ugh
a 50
Ω
re
si
stor. T
o
me
a
s
ure the
vibration
signal
s by u
s
ing thi
s
accele
rom
e
te
r a
nd
analy
s
i
s
of
vibratio
n
in differ
ent
co
ndition. T
he i
nput
sign
al o
n
ea
ch
chan
n
e
l
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 14, No. 3, June 20
15 : 434 – 44
0
436
is co
ndition
e
d
and b
u
ffered then
sam
p
led by a
24
-bit Delta
-
Sig
m
a ADC. T
o
minimize
ground
noise, make
sure the cha
s
sis g
r
ou
nd is
con
n
e
c
ted to earth g
r
ou
nd.
Each chan
n
e
l has p
r
ote
c
t
e
d
from over volt
age
s.
The NI 922
7 C Serie
s
cu
rrent input mo
dul
e wa
s de
signed to mea
s
ure 5 Arms
nominal
and up to 14
A peaks on e
a
ch
cha
nnel
with ch
ann
el-to-ch
ann
el isolation.
Figure 2. BLDC moto
r wit
h
vibration se
nso
r
usi
ng La
b VIEW
4. Block Dia
g
ram
of BL
D
C
M Driv
e
The bl
ock
di
agra
m
of BL
DC Moto
r Dr
ive sh
own
in
the Fig
u
re
3. The
3
Ф
, 400V, AC
sup
p
ly is fed
to Intelligent
Powe
r Mo
dul
e (IPM) th
rou
gh a
u
to transformer.
It co
n
s
ist
s
of
re
ctifier
and
inve
rter circuits. Re
cti
f
ier
is
convert
ed
AC
into DC supply and
inverter is co
nverted
DC i
n
to
AC. DC generator
was coupled with
BLDC
shaft, it act as a eddy cu
rrent load. Hall-effect current
and voltag
e
sen
s
o
r
s
are
use
d
to mea
s
ure the
volt
age a
nd
current
signal
s.
The torque
is
measured by
usin
g load
cell; it will indi
cate t
he to
rq
ue value
s
in
torque i
ndi
ca
tor. The FPG
A
controlle
r giv
e
s
prope
r
co
mmutation to
the i
n
verte
r
c
i
rc
uit via buffer IC’s
. The ac
tive filter i
s
con
n
e
c
ted b
e
twee
n rectifi
e
r output an
d
invert
er
i
n
put. An a
c
ce
lerom
e
ter is
mounted
on
th
e
BLDCM, it was
con
n
e
c
te
d throu
gh th
e NI cDAQ
card. The
re
al time vibra
t
ion sig
nals
are
measured using Lab VIEW software.
The whol
e
drive syste
m
is controll
ed by Spartan-6
pro
c
e
s
sor. Th
e Block di
ag
ram of BLDCM drive is sh
own in Fig
u
re
3.
Figure 3. Block
Diag
ram o
f
BLDC Moto
r Drive
Each
ph
ase
of voltage
a
nd
cu
rre
nt to
be
sen
s
ed
i
n
Hall Effect
voltage
and
cu
rrent
sen
s
o
r
mod
u
l
e
and spee
d sign
als al
so g
e
t in from Intelligent Powe
r Modul
e. Torque indi
cato
r is
use
d
in
this
work for indi
cate
the
BLDC motor torque
wh
en
addi
ng
load
to th
e
motor
and
fe
d to
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Wa
velet Tran
sform
Based
Fault Diag
no
sis
of BL
DC
Motor Drive (Ram
esh Bal
a
ji.S.M)
437
Hall Effect voltage and
current sen
s
or
module. Data
Acqui
sition System is used to acq
u
ire
rea
l
time cu
rrent, voltage, sp
ee
d, torqu
e
of the BL
DC m
o
tor throug
h Hall Effect voltage a
nd
cu
rrent
sen
s
o
r
mod
u
l
e. Hall sen
s
or is u
s
ed t
o
sen
s
e
s
th
e rotor p
o
siti
on, whi
c
h is given to FPG
A
controlle
r. These Spa
r
tan
-
6 FPGA enhanced and
co
mbined
with advan
ced p
r
o
c
e
ss te
chnol
ogy
deliver mo
re
functionality
and ban
dwidth per
doll
a
r than
wa
s previou
s
ly possibl
e. The
accele
rom
e
te
r is u
s
e
d
to
perfo
rm the
measurem
ent
of the vibrati
on si
gnal
s a
nd u
s
ing
cu
rrent
DAQ
NI 92
27 to m
e
a
s
ure th
e
cu
rrent
sign
als by La
b VIEW
softwa
r
e. It is pl
a
c
ed
circumfe
renti
a
lly on the BLDC moto
r bo
dy.
5. Experimental Re
sult a
nd Discu
ssi
on
The experim
ent setup for BLDC m
o
to
r drive through SPART
A
N-6
FPGA
kit and
vibration
sen
s
or th
rou
gh L
ab VIEW is shown in
the F
i
gure
4. The
developm
ent control sy
ste
m
is
tested
on
a F
P
GA ba
sed
BLDC Moto
r
drive
setup
in
Electri
c
al
Dri
v
es a
nd
Co
ntrol la
bo
ratory
at
K.S.Ranga
sa
my College of
Techn
o
logy.
Figure 4. Hardwa
re Setup
of BLDC mot
o
r
5.1. Time and Frequen
c
y
Featur
e of V
i
bration Sign
als during Health
y
Condition
The di
screte
wavelet tran
sform featu
r
e
s
of
vibration
sign
als are
shown in
Figu
re 5
an
d
Figure 6 sho
w
s m
a
ximum,
minimum, m
ean an
d sta
n
dard
deviatio
n
sign
als
(SDS) for the h
e
a
l
th
y
pha
se of vibration of the BLDC motor i
s
ta
ken u
s
in
g the discrete wavelet transfo
rms.
Figure 5. DWT Vibration Signal
s durin
g He
alth
y Condition
%
D
R
=
st
d
(
D2
)+
st
d(
D3
)+
st
d
(
D4
)
+
st
d
(
D
5
)
+
st
d
(
D
6
)/
st
d(
D1)
*
10
0
(1)
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 14, No. 3, June 20
15 : 434 – 44
0
438
Figure 6. DWT Mean, Maxi
mum, Minimu
m, SDS of th
e Healthy Co
ndition
The de
comp
osition value
is upto 8 leve
ls and moth
e
r
wavelet u
s
e
d
is “db
2
” si
g
nal. The
curre
n
t value
in th
e first F
i
gure
i
s
h
ealt
h
y and
the
d
e
com
p
o
s
ition
ratio
is ab
ou
t 12.627
5. T
he
time an
d fre
quen
cy featu
r
es a
r
e
extra
c
ted i
n
th
e
MATLAB u
s
i
ng the
mat
file p
r
og
ram
i
n
the
comm
and
wi
ndo
w in whi
c
h is used find
the fault val
ues of the time
and freq
uen
cy of vibratio
n in
BLDC m
o
tor.
5.2.
Time and Frequen
c
y
Featur
e duri
ng Open
Circuit Fault
Co
ndition
The extra
c
te
d time and frequ
en
cy feature
usi
ng di
screte wavel
e
t transfo
rm
s during
singl
e pha
se
open ci
rcuit fault condition is sh
own
in Figure 7.
The fault dia
gno
sis vibration,
mean, maxi
mum, minim
u
m, stan
da
rd
deviation
si
gnal
s (S
DS)
for op
en
circuit fault vibra
t
ion
sign
als of BL
DC m
o
tor is
taken u
s
in
g the discr
ete
wavelet tra
n
sform sh
own in Figure 8. The
level of deco
m
positio
n rati
o durin
g the o
pen ci
rcuit fault is about 3
4
.9682.
Figure 7. DWT Vibration Si
gnal
s duri
ng
Faulty Conditi
on
%
D
R
=
st
d
(
D2
)+
st
d(
D3
)+
st
d
(
D4
)
+
st
d
(
D
5
)
+
st
d
(
D
6
)/
st
d(
D1)
*
10
0
(2)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Wa
velet Tran
sform
Based
Fault Diag
no
sis
of BL
DC
Motor Drive (Ram
esh Bal
a
ji.S.M)
439
Figure 8. DWT Mean, Maxi
mum, Minimu
m, SDS of th
e Faulty Con
d
ition
Table 1 give
s the details a
bout the extracted
featu
r
e
s
of the BLDC motor vibration for
variou
s op
erating conditio
n
s. Th
e mea
n
, minimu
m,
maximum, st
anda
rd d
e
via
t
ion value d
u
r
ing
the he
althy condition
an
d
open
ci
rcuit f
ault conditi
o
n
are give
n a
n
d
it is u
s
eful
i
n
an
alyzin
g t
h
e
vibration si
gn
al whi
c
h give
s the inform
ati
on abo
ut fault type and the time of fault.
Table 1. Extracted F
eatures for
Differe
nt
Operating Con
d
ition of the BLDC Mot
o
r
Data set
Extract
features
Wavelet sub band (Hz)
D1
D2 D3 D4 D5
D6
D7 D8
Health
y
Condition
Maximum
0.0058
0.0019
0.0033
0.0031
8.3e-04
5.7e-04
4.1e-04
1.7e-04
Minimum
-0.006
-0.002
-0.003
-0.003
-9.7e-0
4
-0.0011
-4.4e-0
4
-1.6e-0
4
Mean
4.5e-08
1.1e-09
-5.4e-0
7
-1.8e-0
7
4.3e-07
1.9e-07
-8.3e-0
7
-9.1e-0
7
Standard
Deviation
0.0022
4e-04
5.7e-04
5.8e-04
1.8e-04
2.5e-04
1.4e-04
3.4e-05
Decompos
ition Ratio
12.6277
Open
Circuit
Fault
Condition
Maximum
0.0036
0.0018
0.0033
0.0029
5.1e-04
5.9e-04
0.0012
9.8e-04
Minimum
-0.0032
-0.0019
-0.0028
-0.0023
-5.3e-0
4
-0.0012
-0.001
-8.2e-0
4
Mean
1.3e-06
-1.6e-0
6
2.2e-06
5.9e-06
5.9e-07
-9.6e-0
6
-3.6e-
06
4.1e-06
Standard
Deviation
8.9e-04
3.9e-04
8.3e-04
5.6e-04
2.6e-04
2.4e-04
3.1e-04
3.3e-04
Decompos
ition Ratio
34.9682
The freque
ncy sub
ban
d value give
s th
e inform
ation
about th
e hi
gher orde
r freque
ncy,
fundame
n
tal
freque
ncy
and the lo
wer orde
r fre
quen
cy valu
es in
whi
c
h
the fundam
ental
freque
ncy lie
s at the sixth level of deco
m
positio
n.
6. Conclusio
n
In this pa
pe
r reviewed
si
ngle p
h
a
s
e o
pen
circuit fa
ult con
d
ition
vibration a
n
a
l
ysis of
BLDC M
o
tor and extra
c
t the signal f
eature
s
fro
m
the Discrete Wavelet T
r
an
sform
(DWT).
FPGAs ba
se
d
on
spe
ed control of
thi
s
system
re
veal the
u
s
e
of reliabl
e
and l
o
wer cost
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 14, No. 3, June 20
15 : 434 – 44
0
440
controlle
r. Th
e sp
eed
of BLDCM is con
t
rolled by
dut
y cycle a
nd freque
ncy of t
he si
mple Pu
lse
Width M
odul
ation (P
WM
) tech
nique
s.
In this a
naly
s
is sin
u
soida
l
PWM
cont
rol tech
niqu
es is
employed.
T
he p
r
op
osed
syste
m
i
s
analyzed
th
e
vibration
in
fault conditi
on a
nd
heal
th
y
con
d
ition
with
spe
ed of 3
0
00rp
m
. The
result
s are obt
ained
by usi
n
g Lab VIEW
analysi
s
of b
o
th
time domain
and freq
uen
cy domain. Th
e analyzed si
gnal
s are extracted in
DWT
using MATL
AB,
unde
r the fa
ulty condition
vibration to
be diag
no
si
s by extractin
g
sign
al feat
ure
s
by wavelet
transfo
rm to be analyzed
in values of maximum, minimum, mea
n
and stan
da
rd deviation.
In
these inve
stig
ations
we ha
ve to find the distorti
o
n
rati
on of both he
althy conditio
n
(12.6
277
) a
nd
faulty conditio
n
(34.96
82
).
Referen
ces
[1]
A
y
h
an, MY Ch
o
w
, MH S
ong.
Multipl
e
discri
m
inate a
nal
ys
i
s
and ne
ura
l
-n
et
w
o
rk-b
ased
mono
lith an
d
partitio
n
fault-d
e
tection sc
hem
es for broke
n
r
o
tor bar i
n
in
d
u
ction mot
o
rs.
IEEE Trans. Ind. Electron
.
200
6; 53(4): 12
98-1
308.
[2]
Basz
ynski. A n
o
vel sp
eed me
asurem
ent method for a hi
gh-
spee
d BLDC
motor base
d
o
n
the sign
al
s
from the rotor positi
on se
nsor
.
IEEE
Transac
tions on Indust
r
ial Infor
m
atics
.
2013: 2
31-2
4
0
.
[3]
B
y
o
u
n
g
-Gun
Park. F
ault T
o
ler
ant Strate
gies f
o
r BLD
C
Motor Dr
ives
un
der S
w
itc
h
F
aults.
IEEE
T
r
ansactio
n
.
2006: 16
37-
164
1.
[4]
Camero
n, W
T
T
homson, AB Do
w
.
Vi
bratio
n and
c
u
rrent mo
nitori
ng
for det
ecting air-g
ap eccentric
i
ty
in
larg
e ind
u
ctio
n motors.
Proc. Inst. Electr. Eng. 1985; 13
3(3): 155-
163.
[5]
Gunja
n
Sarda
na, Nee
l
am Meha
la.
F
ault
Identificatio
n
of BLDC Mo
tor Using Sig
nal Proc
essin
g
T
e
chni
ques
−
A Review
. Natio
n
a
l Co
nferenc
e
of Science i
n
M
edi
a, ISSN (Onlin
e). 201
2: 227
7-96
98.
[6]
Khomfoi. LM T
o
lb
ert. F
ault di
agn
osis a
nd r
e
config
uratio
n f
o
r multil
eve
l
in
ve
rter driv
e usi
ng AI-bas
e
d
techni
qu
es.
IEEE Trans. Ind.
Electron
. 20
07;
54(6): 295
4-2
968.
[7]
Kirti B Satale. Analy
s
is of Different Speed C
ontrol
l
ers an
d Impleme
n
tatio
n
of Novel Spe
e
d
Contro
lle
r
using FPGA for BLDC Motor.
Internati
ona
l
Journ
a
l
of Soft Co
mp
utin
g a
n
d
En
gin
eeri
ng.
201
4; 4(
2):
131-
134.
[8]
Murat YILMAZ
. Sensorless
Drive of Brush
l
e
ss DC Motor
Based up
on
W
a
velet Ana
l
ysis. Istanbul
T
e
chnical Un
iv
ersit
y
. 2
005
[9]
Madh
urima
C
h
attopad
h
y
a
y
.
S
i
mulati
on Mo
de
ling of BLD
C
motor driv
e an
d har
mo
nic an
aly
s
is of stator
current, rotor
spe
e
d
an
d
accel
e
ratio
n
u
s
ing
Discr
ete
W
a
vel
e
t T
r
a
n
sform T
e
ch
ni
que.
Pr
ocedia
T
e
chnolog
y. 2
012; 4: 66
6-67
0.
[10]
Pind
ori
y
a.
Spe
ed Co
ntrol of BLDC Motor u
s
ing Sin
u
so
ida
l
PW
M
T
e
chniq
u
e
. Natio
nal C
onfere
n
ce o
n
Emergi
ng T
r
ends in Com
puter
& Electrical En
gin
eeri
ng. 20
1
4
: 151-1
55.
[11]
Sath
ya. Ana
l
ys
is and
e
x
per
im
ental i
n
vesti
gat
ion to
miti
gate
mecha
n
ica
l
vib
r
ation a
nd ac
o
u
stic nois
e
in
direct torqu
e
c
ontrol
l
er fed P
M
SM.
Internationa
l Jour
nal of
Researc
h
in E
ngi
neer
in
g an
d
T
e
chnol
ogy
.
201
4: 232
1-73
08.
[12]
Satish R
a
ja
go
pal
an. An
al
ytic
-W
avelet-Ri
d
g
e
-Bas
e
d
Det
e
ction of D
y
n
a
mi
c Eccentricit
y
i
n
Brush
l
e
s
s
Direct Curr
ent
(BLDC) Motors
F
unction
i
ng U
nder D
y
nam
ic Operatin
g
Co
n
d
itio
ns.
IEEE Transactions
on Industri
a
l El
ectronics.
20
07
; 54(3): 141
0-1
419.
[13]
Sh
yu, Y Ch
an
g. Nove
l loss r
educti
on
puls
e
w
i
dth m
odu
lati
on tech
ni
que f
o
r brus
hless
d
c
motor driv
e
s
fed b
y
MOSF
E
T
inverter.
IEEE Trans. Power Electron
. 200
4; 19(6): 16
46-
165
2.
[14]
T
aeyon
g Yoo
n
.
Magnetica
l
l
y
Induced Vi
br
ation i
n
a Per
m
ane
nt-Magn
e
t
Brushless D
C
Motor
w
i
t
h
S
y
mmetric Po
l
e
-Slot Co
nfigur
ation.
IEEE Transactions on M
agnetics.
200
5; 41(5).
[15]
T
hati Venkata,
M Laks
h
mi.
W
a
velet A
nal
ysis Ba
se
d Ove
r
current Pr
ote
c
tion for P
e
rm
ane
nt Mag
net
Brushl
ess DC
Motor.
Internati
ona
l Journ
a
l of
Soft Comp
utin
g and En
gi
neer
ing.
20
13; 3(4):
2231-
23
07.
Evaluation Warning : The document was created with Spire.PDF for Python.