Escolar Documentos
Profissional Documentos
Cultura Documentos
Received 20 April 2004; received in revised form 10 August 2004; accepted 18 August 2004
Available online 4 May 2005
Abstract
This paper presents two orthogonal axis models for simulation of three-phase induction motors having asymmetrical windings and inter-turn
short circuits on the stator. The first model assumes that each stator phase winding has a different number of turns. To model shorted stator
turns, the second model assumes phase as has two windings in series, representing the unaffected portion and the shorted portion. It uses the
results of the first model to transfer phase as to qd so that shorted portion is transferred to the q axis. Simulations results from the models are
in good agreement with other studies and are compared with experiment carried out on a specially wound motor with taps to allow different
number of turns to be shorted. The models have been successfully used to study the transient and steady state behaviour of the induction motor
with short-circuited turns, and to test stator fault diagnostic algorithms operating in real time.
2005 Elsevier B.V. All rights reserved.
Keywords: Induction motors; Turn faults; Modelling; Diagnostics; Orthaogonal axis
1. Introduction
Because of costly machinery repair, extended process
down time, and health and safety problems, a trend in modern industry is to focus attention and resources on fault detection and predictive maintenance strategies for industrial
plant [1,2]. It is known that approximately 36% of induction
motor failures are caused by failure of the stator winding, and
it is believed that these faults begin as undetected turn-to-turn
faults in a coil, which progress to catastrophic phase-to-phase
or phase-to-ground short circuit faults [1,2]. To achieve prior
warning of failure so that an orderly shut-down may be made
to avoid catastrophic failure, shorted turns within a stator
winding coil must be detected or predicted [15].
Corresponding author.
E-mail addresses: markan@inonu.edu.tr (M. Arkan),
dragica.kostic-perovic@dana.com (D. Kostic-Perovic),
p.j.unsworth@sussex.ac.uk (P.J. Unsworth).
0378-7796/$ see front matter 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.epsr.2004.08.015
58
to qd0 axis. The same transformation process can be applied to machines in which there are phase unbalances [13].
Hence, it is useful to extend this approach to also incorporate problems encountered with asymmetrical induction
motors.
The aim of this paper is to present a useful and straightforward method to simulate inter-turn short circuits for
diagnostic purposes. The fault can be simulated by disconnecting one or more turns making up a stator phase winding
[9,14]. Firstly an induction motor model with unequal
numbers of stator turns has been developed. Then using this
model, a second model has been developed to simulate stator
inter-turn short circuits. Models are simulated in Matlab
Simulink and simulation results are presented. Results
obtained are confirmed with a conventional asymmetrical
motor model in a three-phase, non-orthogonal base, and
by experimental results obtained from a specially wound
motor.
s
r11
s
r12
s
r13
s
s
s
r22
r23
where rsqd0 = r21
and matrix elements
s
s
s
r31
r32
r33
are given in Appendix A and assuming rar = rbr = rcr =
rr , rrqd0 = rr I3x3 .
In matrix notation, the flux linkages of the stator and rotor
windings may be written in terms of the winding inductances
and the current as
sr
sabc
Lss
isabc
abc Labc
=
(3)
rr
rabc
Lrs
irabc
abc Labc
where stator
and
rotor
inductances
are
Lss
and
Lrr
abc =
abc = Lbsas Lbsbs Lbscs ,
Lcsas Lcsbs Lcscs
2
2
L
cos
L
cos
cos
L
asar
r
ascr
r
r
asbr
3
3
2
2
sr
(4)
Labc = Lbsar cos r
Lbscr cos r +
Lbsbr cos r
3
3
2
2
Lcsbr cos r
Lcscr cos r
Lcsar cos r +
3
3
With the appropriate subscripts as, bs, cs, ar, br, and cr,
the voltage equations of the magnetically coupled stator and
rotor circuits can be written as follows:
sr
and Lrs
abc = Labc where ( ) means the transpose of the matrix.
The coefficients Lasar , Lasbr , Lascr , Lbsar , Lbsbr , Lbscr , Lcsar ,
Lcsbr , and Lcscr are peak values of stator-to-rotor mutual inductances. Because of rotor symmetry Lasar = Lasbr = Lascr ,
Lbsar = Lbsbr = Lcscr , and Lcsar = Lcsbr = Lcscr .
The stator and rotor qd0 flux linkages are obtained by applying transformation to the stator and rotor abc flux linkages
in Eq. (3), that is
s
sr r
sqd0 = Lss
qd0 iqd0 + Lqd0 iqd0 ,
(1)
where p = d/dt. Applying a stationary reference frame transformation to this equation yields the corresponding qd0 equations and Eq. (1) becomes
vsqd0
rsqd0 isqd0
+ psqd0 ,
0 = rrqd0 irqd0 r 1
0
1
0
0
0 rqd0 + prqd0
0
s
rr r
rqd0 = Lrs
qd0 iqd0 + Lqd0 iqd0
ss
L11
ss
ss
where
Lqd0 = L21
(2)
Lsr
11
sr
L21
Lsr
31
Lsr
12
Lsr
22
Lsr
32
0 ,
0
Lss
31
Lrr
qd0
(5)
Lss
12
Lss
13
Lss
32
Lss
33
Lss
22
Lrr
11
= 0
0
Lss
23 ,
0
Lrr
22
0
Lsr
qd0 =
0
0 ,
Lrr
33
and
0.5 Lsr
31
sr
0.5 Lsr
Lrs
32 . Matrix elements of
qd0 = L12
0
0
0
sr
rr
rs
,
L
Lss
,
L
,
and
L
qd0
qd0
qd0
qd0 are given in Appendix A.
Normally, an induction machine is connected to a threephase supply by a three-wire connection (i.e. neutral current
does not flow). Hence, for a squirrel cage induction machine
and three-wire connection, the stator and rotor flux linkages
in Eq. (5) may be expressed compactly as
sq
Lsr
11
Lsr
21
Lsr
22
Lss
11
s Lss
d 21
r = sr
q L11
Lsr
21
rd
Lss
12
Lsr
11
Lsr
12
Lrr
11
0
Lss
22
Lsr
22
Lsr
21
Lsr
12
isq
s
Lsr
22 id
0 irq
Lrr
22
(6)
ird
N2
= a2
Ns
2
Lls + Lm
3
Na2 Lmls
(9)
2 Nr2
Lm = Llr + Lmar (13)
3 Ns2
N2
1
2 Nr2
1
= Lcrbr =
L
m = Lmar
2
3 Ns2
2
(14)
2 N a Nr
Lm = Na Lmsr
3 Ns2
(15)
2 Nb N r
Lm = Nb Lmsr
3 Ns2
(16)
2 Nc N r
Lm = Nc Lmsr
3 Ns2
(17)
where Lmsr =
(8)
2 Nr
3 Ns2 Lm .
where Lmls =
Lls +
.
The stator mutual inductances between phases as and bs,
bs and cs, and cs and as can be derived as
1 N a Nb
2 Lm
1
=
Lm
Lasbs = Lbsas = Na Nb
2
3 Ns 2
3 Ns2
1
Ns2
Therefore
(7)
59
2
3 Lm
= Na Nb Lmss
In this subsection, equations are rearranged for the asymmetrical induction motor model that has been developed.
The qd-applied voltage in a reference frame fixed to the
stator can be obtained from the stator phase voltage vas , vbs ,
and vcs by standard transformation [17]. The qd voltages will
be
(10)
(11)
(12)
60
2
1
1
2
vas (vbs + vcs ) =
vag (vbg + vcg ) ,
3
2
3
2
vsq =
1
1
vsd = (vbs + vcs ) = (vbg + vcg )
3
3
(18)
s s
s s
(vsd r21
iq r22
id ) dt,
(r rq + rrr ird ) dt
(19)
2 Lm
2
Lls + Lm + 2Nus Nsh
3
3 Ns2
N2
2
Lls + Lm
+ sh
2
3
Ns
N2
= us
Ns2
2
= (Nus
Lmls + Nus Nsh Lmsh )
2
+(Nus Nsh Lmsh + Nsh
Lmls )
2 Lm
3 Ns2
Lbsbs = Lcscs =
Ns2
Ns2
(22)
2
Lls + Lm
3
= Ns2 Lmls
(23)
1 N a Ns
1 Nus
1 Nsh
Lm =
Lm
Lm
3 Ns2
3 Ns
3 Ns
= Lasbs + Lshbs
(24)
(25)
2 N a Nr
Lm
3 Ns2
61
2 Nus Nr
2 Nsh Nr
Lm +
Lm = Lasar + Lshar
3 Ns2
3 Ns2
(26)
2 Ns N r
Lm
3 Ns2
(27)
Lss
qd0
Lss
11
= 0
0
Lss
22
Lss
31
Lsr
11
0
Lsr
22
Lsr
qd0 = 0
Lsr
31
Lss
13
0 and
Lss
33
0
0
Lrs
qd0
Lrr
33
Lsr
11
= 0
0
0
Lsr
22
0.5Lsr
31
0
0
(28)
sd
(29)
sh
q
s
q
s
d
r
q
rd
Lsh
q
ssh
Lq
= 0
shr
Lq
0
Lssh
q
Lsq
0
Lsr
q
0
0
0
Lsd
0
Lsr
d
Lshr
q
Lsr
q
0
Lrq
0
ish
q
0 isq
s
Lsr
d id
0 irq
Lrd
(30)
ird
Na
Nus
Nsh
rs =
rs +
rs = ras
+ rsh
Ns
Ns
Ns
rbs = rcs = rs
(31)
(32)
s
r21
s
r22
rr
r
r
where Lrr
11 = Lq and L22 = Ld . Matrix elements are given
in Appendix A.
The stator and rotor flux linkages for the new model of a
squirrel cage induction motors will be
where rsh is the shorted winding(s) resistance. Stator qd resistance for new model will be
sh 2
rq
0
3 rsh 0
s
s
s
(33)
r11
r12
rq = 0
rds
ss
s
sr
s
ssh
sh
ssh
where Lss
11 = (Lq + Lq ) + (Lq + Lq ), L22 = Ld , L11
sr
sr
sr
shr
= Lq + Lq and L22 = Ld .
Similarly, the results of transformation of rotor self- and
mutual inductances can be simplified as
rr
0
0
L11
Lrr
0 and
Lrr
qd0 = 0
22
rq =
rd =
s s
s s
(vsq vsh
q r11 iq r12 id ) dt,
s s
s s
(vsd r21
iq r22
id ) dt,
(r rq + rrr ird ) dt
(34)
4. Simulation results
The models developed have been simulated in Matlab
Simulink and the following simulation results are for a 2 hp
motor whose parameters are given in Appendix A. The results
are taken during acceleration from stand still to full speed.
Figs. 4 and 5 are results for acceleration from stand still
to full speed at full load under normal conditions. Initially,
motor currents are not symmetrical because of the starting
62
transient. This introduces high negative sequence current until the motor reaches full speed.
Figs. 6 and 7 show the simulation results from stand still
to full speed at full load with five turns shorted. The supply
negative sequence current is around 400 mA. Because of stator asymmetry, the torque graph in Fig. 6 shows the expected
pulsation at twice supply frequency (2fs ) even at steady state.
This is because the negative sequence current introduces a
braking torque in the motor. The short circuit current of 60 A
in the shorted turns seen in Fig. 7 is 2.6 times the normal
locked rotor current of 23 A.
Figs. 8 and 9 are for same condition but with an external
1.5 resistance to limit short circuit current to avoid destruction of the motor. This limits short circuit current to nearly
3 A, and supply negative sequence current to 27 mA. The
torque still has pulsation 2fs but is much smaller as seen in
Fig. 8.
Fig. 7. Positive sequence, negative sequence, and short circuit currents for
five turns shorted.
Fig. 8. Torque and speed variation with five turns shorted with 1.5 resistance.
Fig. 9. Positive sequence, negative sequence and short circuit currents for
five turns shorted with 1.5 resistance.
Results have also been obtained by simulation for different numbers of directly shorted turns (rext = 0). For rext = 0
negative sequence current In is proportional to the number n
of shorted turns
In
= kn
Fig. 10. Short circuit current vs. shorted turns for different external resistance values (rext ).
In n or
63
If a motor is operated from an unbalanced voltage supply, the heating effect of unequal phase currents will cause
(35)
where k is constant.
Fig. 10 shows Ir ext current in the external resistance and
number of shorted turns relation with 0 and 0.2 external
current limiting resistance. As it can be seen from figure short
Table 1
Experimental and simulation results for different numbers of shorted turns
with short circuit current limited externally
n
1
2
3
4
Experimental result
Simulation result
In (mA)
In (mA)
Ir ext (A)
5.6
16
31.4
48.5
2.9
5.6
8.1
10.2
4
15
30
54
Ir ext (A)
2.7
5.3
8.98
10
Fig. 11. Negative sequence current vs. shorted turns for different external
resistance values (rext ). Short circuit current limited by 0, 0.1, 0.2 and 0.5
resistance.
64
(41)
Fig. 12. Negative sequence current vs. extra series resistance runb .
(36)
Isp
runb
3 Zn + (runb /3)
(37)
where
Isp is positive
sequence current, Zn =
r
rr
the above
rs + 2sp + j(Xs + Xr ), and if Zn unb
3
equation can be simplified to
In =
runb Isp
3 Zn
(38)
3 n Vsp
8 N s Zs
(39)
(40)
5. Conclusion
Two models have been developed for analysing an
asymmetrical induction motor. The models are based on
general machine parameters so that it is not necessary to
know detailed motor geometry or physical layout of the
windings. New parameters for asymmetrical conditions
have been presented. The models have also been confirmed
against simulation results obtained from a conventional
asymmetrical motor model in a three phase non-orthogonal
base, which requires seven differential equations. It has not
been presented because of space.
Inter-turn faults can be easily simulated by the model.
Fault severity can be controlled by the number of shorted
turns and an optional current limiting resistance to short
circuit the windings. Simulation results have been used in
[2,4,20,21] to check fault detection algorithms for stator
faults.
The models have been used to study the relation between
the number of shorted turns and negative sequence current,
the effects of resistive unbalance between the phases due to
the heating effect of unbalanced phase currents, and temperature effects on shorted turn resistance. Results confirm that
negative sequence current and shorted number of turns are
linearly dependant, and that resistive unbalance can produce
comparable negative sequence current, so that fault detection
algorithms should be able to distinguish between the effects
of turn shorts and thermal unbalance [2,4,20]. The results obtained for short circuit under limited short circuit current (by
external resistance) reveal that sensitivity depends on number
of shorted turns and short circuit current.
Acknowledgements
The authors thank Dr. Peter Lindon for providing a conventional asymmetrical motor model in a three-phase nonorthogonal base to validate the developed models, and for his
valuable discussions.
ss
ss
ss
(7)(13) it can be shown that Lss
12 = L13 = L21 = L23 =
ss
ss
ss
ss
ss
L31 = L32 = 0 and L11 = L22 = Lls + Lm , L33 = Lls .
Appendix A
A.1. Matrix elements for induction motor with different
numbers of stator turns
The qd0 resistances elements are:
2
1
1
s
r11 =
ras + rbs + rcs
3
4
4
3
s
=
r12
(rbs rcs )
6
(42)
(43)
s
r13
=
1
(2ras rbs rcs )
3
(44)
s
r22
=
1
(rbs + rcs )
2
(45)
1
= (ras + rbs + rcs )
3
s
r33
(46)
1
Lss
12 = (Lbsbs Lcscs Lasbs + Lascs )
2 3
Lss
13 =
2
(Lasas .5Lbsbs .5Lcscs + .5Lasbs
3
+ .5Lascs Lbscs )
1
Lss
21 = (.5Lbsbs .5Lcscs Lasbs + Lascs )
3
(55)
(56)
(57)
sr
sr
sr
Lsr
21 = L12 , and L32 = L12 . When Na = Nb = Nc = Ns ,
Lasar = Lasbr = Lascr = Lbsar = Lbsbr = Lbscr = Lcsar = Lcsbr = Lcscr
Nr
sr
and mutual inductances will be Lsr
11 = L22 = Ns Lm and
sr
sr
sr
sr
L12 = L21 = L31 = L32 = 0.
3
Nr2
rr
Lrr
=
L
=
L
+
=
L
+
Lm
L
mar
lr
lr
11
22
2
Ns2
(58)
Lrr
33 = Llr
(59)
(48)
(49)
(50)
1
= (Lbsbs + Lcscs 2Lbscs )
2
(51)
Lss
23
1
= (Lbsbs + Lcscs Lasbs + Lascs )
3
(52)
1
(Lasas + Lbsbs + Lcscs + 2Lasbs
3
+ 2Lascs + 2Lbscs )
(54)
(47)
Lss
22
Lss
33 =
Lsr
11 = Lasr + .25Lbsr + .25Lcsr
3
sr
(Lbsr Lcsr )
L12 =
4
3
Lsr
22 = (Lbsr + Lcsr )
4
Lss
qd0 matrix elements are:
2
(Lasas + .25Lbsbs + .25Lcscs Lasbs
3
Lascs + .5Lbscs )
Lsr
31 = .5Lasr .25Lbsr .25Lcsr
s = r s , r s = 1 r s , r s = 1 r s , and r s = r s , where
r21
12 23
32
12
2 12 31
2 13
Nb
Nc
a
ras = N
r
,
r
=
r
,
and
r
=
r
,
when
N
= Nb =
s
s
cs
s
a
bs
Ns
Ns
N
s = rs =
s = r and r s s = r s = r s = r s =
Nc = Ns , r11
r33
s
22
12
13
21
23
s = r s = 0.
r31
32
Lss
11 =
65
sh
ssh
= (Lsq + Lssh
q ) + (Lq + Lq )
(60)
1
(Lbsbs + Lcscs 2Lbscs ) = Lbsbs Lbscs
2
= Lls + Lm = Lsd
(61)
Lss
22 =
(53)
1 ss
1 ss
ss
and Lss
31 = 2 L13 , and L32 = 2 L23 . When Na = Nb = Nc ,
Nasas = Nbsbs = Ncscs and Nasbs = Nascs = Nbscs . By using Eqs.
(62)
66
3
3
(63)
(Lbsr + Lcsr ) = Lbsr = Lsr
d
4
2
The rotor self- and mutual inductances and resistances will
be the same as before.
Lsr
22 =
460 V
2.7 A
Power factor
Service factor
0.815
1.15
2 hp
1752 rpm
252
4.05
13.97 mH
2.6
13.97 mH
538.68 mH
0.06 kg m2
References
[1] W.T. Thomson, A review of on-line condition monitoring techniques
for three-phase squirrel-cage induction motors past present and future, in: Proceedings of the 1999 IEEE International Symposium on
Diagnostics for Electrical Machines, Power Electronics and Drives,
IEEE SDEMPED99, 1999, pp. 318.
[2] M. Arkan, Stator Fault Diagnosis in Induction Motors, University of
Sussex, Ph.D. Thesis, 2000.
[3] R.M. Tallam, T.G. Habetler, R.G. Harley, Transient Model for Induction Machines with Stator Winding Turn Faults, in: Proceedings of
the IEEE Industry Applications Conference 35th Annual MeetingWorld Conference on Industrial Applications of Electrical Energy,
Rome, Italy, October 2000, 2000.
[4] M. Arkan, D. Kostic-Perovic, P.J. Unsworth, Online stator fault diagnosis in induction motors, IEE Proceedings: Electric Power Applications 148 (6) (2001) 537547.
[5] G.B. Kliman, W.J. Premerlani, R.A. Koegl, D. Hoeweler, A New
Approach to On-Line Turn Fault Detection in AC Motors, in: IAS
Annual Meeting, October 1996, 1996, pp. 687693.
[6] J.E. Brown, O.I. Butler, A general method of analysis of three phase
induction motors with asymmetrical primary connections, Proc. IEE
100 (1953) 2534.
[7] C.S. Jha, S.S. Murthy, Generalised rotating-field theory of woundrotor induction machines having asymmetry in stator and/or rotor
windings, Proc. IEE 120 (8) (1973).
[8] X. Luo, Y. Liao, H.A. Toliyat, T.A. Lipo, Multiple coupled circuit
modelling of induction machines, IEEE Trans. Ind. Appl. 31 (2)
(1995) 311317.
[9] H.A. Toliyat, T.A. Lipo, Transient analysis of cage induction machines under stator, rotor bar end ring faults, IEEE Trans. Energy
Convers. 10 (2) (1995) 241247.
[10] G.M. Joksimovic, J. Penman, The detection of inter-turn short circuits in the stator windings of operating motors, IEEE Trans. Ind.
Electron. 47 (5) (2000) 10781084.
[11] S. Williamson, K. Mirzoian, Analysis of cage induction motors with
stator winding faults, IEEE Trans. Power Apparatus Syst. PAS-104
(7) (1985) 18381842.
[12] A. Consoli, T.A. Lipo, Orthogonal axis models for asymmetrically
connected induction machines, IEEE Trans. Power Apparatus Syst.
PAS-101 (12) (1982) 45184526.
[13] A.K. Wallace, E.S. Ward, A. Wright, Sources of harmonic currents
in slip-ring motors, Proc. IEE 121 (12) (1974).
[14] M.Y. Chow, Methodologies of Using Neural Network and Fuzzy
Logic Technologies for Motor Incipient Fault Detection, World Scientific Publication, 1997, ISBN 981-023265-9.
[15] D.W. Novotny, T.A. Lipo, Vector Control and Dynamics of AC
Drives, Clarendon Press, Oxford, 1996.
[16] P. Krause, O. Wasynczuk, S.D. Sudhoff, Analysis of Electric Machinery, IEEE Inc, 1995, ISBN 0-7803-1101-9.
[17] P. Vas, Vector Control of AC Machines, Oxford Science Publications,
1999.
[18] C.M. Ong, Dynamic Simulation of Electric Machinery, Prentice Hall
PTR, 1998, ISBN 013-723785-5.
[19] S.P. Huang, Active Mains Supply Harmonic Filtering, University of
Sussex, Ph.D. Thesis, 1997.
[20] M. Arkan, P.J. Unsworth, Stator fault diagnosis in induction motors
using power decomposition, in: Phoenix, USA, October, Proceedings
of the IEEE Industry Applications Conference 34th Annual Meeting,
1999, vol. 3, pp. 19081912.
[21] D. Kostic-Perovic, M. Arkan, P.J. Unsworth, Induction motor fault
detection by space vector angular fluctuation, in: Proceedings of
IEEE Industry Applications Conference 35th Annual Meeting-World
Conference on Industrial Applications of Electrical Energy, Rome,
Italy, October, 2000.