IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 38, NO.

1, JANUARY/FEBRUARY 2002 101

Novel Frequency-Domain-Based Technique to Detect

Stator Interturn Faults in Induction Machines Using
Stator-Induced Voltages After Switch-Off
Subhasis Nandi, Member, IEEE, and Hamid A. Toliyat, Senior Member, IEEE

Abstract—Traditionally, for medium- and high-voltage motors 4) contamination due to oil, moisture, and dirt;
and generators, condition-based monitoring of stator faults 5) short circuit or starting stresses;
is performed by measuring partial discharge activities. For
6) electrical discharges;
low-voltage machines, negative-sequence impedance or currents
are measured for the same. Such diagnostic schemes should be 7) leakage in cooling systems.
carefully implemented as supply voltage unbalance, manufac- There are a number of techniques to detect these faults. For
turing-related asymmetry, etc., also produce negative-sequence large generators and motors with stator windings rated 4 kV and
voltages. A few approaches based on motor current signature above, online partial discharge (PD) test methods give very re-
analysis have already been proposed to detect stator interturn
faults. However, little or no physical insight was provided to liable results [3]. Even a portable test instrument called TGA-B
explain the occurrence of certain harmonics in the line current or is available for this purpose.
the influence of voltage unbalance on these harmonics. Also, in at However, for low-voltage motors, stator fault detection pro-
least one of these papers, a large portion of the stator winding was cedures are yet to be standardized. Penman et al. [4] were able to
shorted to emulate the faults. The method proposed in this paper
monitors certain rotor-slot-related harmonics at the terminal detect turn–turn faults by analyzing the axial flux component of
voltage of the machine, once it is switched off. In the absence of the machine using a large coil wound concentrically around the
supply voltage, issues such as voltage unbalance, time harmonics shaft of the machine. Even the fault position could be detected
do not influence the measurements except as initial conditions, by mounting four coils symmetrically in the four quadrants of
which is a very desirable feature when the machine is fed from an
adjustable-speed drive. Satisfactory simulation and experimental the motor at a radius of about half the distance from the shaft
results have been obtained with only about 1.5% (5/324) of the to the stator end winding. The frequency components to be de-
total number of turns shorted. tected in the axial flux component are given by
Index Terms—Condition monitoring, fault diagnosis, induction
machine, stator faults. (1)

where is the number of pole pairs, is the supply frequency,

I. INTRODUCTION and , and is the slip.
Toliyat and Lipo [5] have shown through both modeling and
S TATOR FAULTS are usually related to insulation failure.
In common parlance they are generally known as phase-to-
ground or phase-to-phase faults. It is believed that these faults
experimentation that these faults result in asymmetry in the ma-
chine impedance causing the machine to draw unbalance phase
start as undetected turn-to-turn faults, which finally grow and currents. This is the result of negative-sequence currents flowing
culminate into major ones. Almost 30%–40% of all reported in the line as also has been shown in [6]. However, negative-se-
induction motor failures fall into this category [1]. The primary quence currents can also be caused by voltage unbalance, ma-
causes for failure of stator insulation are as follows [2]: chine saturation, etc. Kliman et al. [1] modeled these unbal-
1) high stator core or winding temperatures; ances. The model also includes instrument asymmetries. It is
2) slack core lamination, slot wedges, and joints; reported that, with these modifications, it is possible even to de-
3) loose bracing for end winding; tect a one turn “bolted” fault out of a total 648 turns. The dif-
ference between the positive-sequence current under the faulty
and the healthy conditions divided by the positive-sequence cur-
Paper IPCSD 01–083, presented at the 2000 Industry Applications Society
Annual Meeting, Rome, Italy, October 8–12, and approved for publication in rent under the healthy condition is also reported to be an effec-
the IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS by the Electric Machines tive diagnostic index [7]. Detection of stator faults using nega-
Committee of the IEEE Industry Applications Society. Manuscript submitted tive-sequence impedance has been addressed in [8]. Statistical
for review May 1, 2000 and released for publication October 29, 2001. This
paper was based in part upon work supported by the Texas Advanced Research process control (SPC) techniques have also been applied to de-
Program under Grant 000512-0079-1999. tect stator faults [9]. Detection of stator voltage unbalance and
S. Nandi is with the Department of Electrical and Computer Engi-
neering, University of Victoria, Victoria, BC V8W 3P6 Canada (e-mail:
single phasing effects using advanced signal processing tech-
snandi@ece.uvic.ca). niques have been described in [10].
H. A. Toliyat is with the Department of Electrical Engineering, Apart from these well known techniques, a few methods
Texas A&M University, College Station, TX 77843-3128 USA (e-mail:
toliyat@ee.tamu.edu). based on motor current signature analysis (MCSA) for interturn
Publisher Item Identifier S 0093-9994(02)00787-9. stator fault detection have been reported recently [11], [12].
0093–9994/02$17.00 © 2002 IEEE
102 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 38, NO. 1, JANUARY/FEBRUARY 2002

In [11], low-frequency components, similar to those observed herent eccentricities also contribute to the visibility of some of
with mixed eccentricity-related faults, are shown to be present these components that should not appear for an ideal machine.
around fundamental and third time harmonics. Motor loading When a few turns of a stator phase are short circuited, they get
attenuated these components. Also, the results were obtained isolated from the balanced three-phase winding set and act as
by shorting a large number of turns (30/361) in one phase. a search coil to all the harmonics described by (3). Thus, the
Faults in stator windings with parallel circuits are discussed short-circuited turns will be able to induce these harmonics in
in [12]. Here, high-frequency components, similar to the ones the healthy portion of the stator winding.
associated with static and dynamic eccentricities, are reported Although MCSA can detect these components, they may be
to have shown changes. However, the physics behind the confused with voltage unbalance in machines. Fortunately, they
existence of such components, particularly the low-frequency can be unambiguously detected at the terminal voltages of the
ones, are not clearly explained. Also, issues such as voltage machine, just after switching it off. Since the supply voltage
unbalance, constructional imperfections which produce similar is absent, voltage unbalance cannot influence the detection
effects, are not addressed. In the proposed novel interturn stator process. Also, since the flux of the machine is predominantly
fault detection scheme, the line–line voltages of the machine, sinusoidal, even when solid-state drives are used, this scheme is
instead of line currents, are monitored for their frequency likely to give good results, even with supply voltages rich with
components, immediately after the machine is switched off. harmonics. In the absence of stator supply, slip of an induction
Normally, the healthy three-phase stator windings pick up motor does not exist. Hence, the frequency components to be
only those harmonics present in the rotor MMF (since there detected in the line–line voltages after switch-off are
is no stator current after switching off the machine) whose
pole pair numbers match with those the stator winding itself (4)
can produce. However, with an interturn short-circuit fault,
the shorted portion of the winding acts as a pickup coil to
which can be obtained by substituting in (3). is the
these harmonics irrespective of their pole pair numbers. As
frequency of the decaying stator voltage after switch-off and is
current still flows in the shorted turns, these harmonics will
proportional to the gradually diminishing speed of the induction
be induced in the healthy portion of the winding. Since the
machine. The time constant of the speed decay is dependent
mutual inductance between the shorted turns and the three
upon the inertia and friction of the drive system. Because of
phases is different, it will also be able to detect the faulty phase
the decreasing machine speed, short-time fast Fourier transform
unequivocally. Also, as the detection takes place with supply
(FFT) has been used to detect these spectral lines. Equation (4)
disconnected, the scheme is largely insensitive to the supply
can also be written as a product of (2) and . Thus, (2) defines
voltage unbalance and time harmonics. Detailed simulation and
the harmonic numbers to check for detecting these faults. Due
experimental results are provided in the following sections.
to the transient nature of the signal processed, detectability may
be good only for .
II. NOVEL STATOR FAULT DETECTION SCHEME
It is known [13]–[15] that the predominant space har- III. MODELING OF INDUCTION MACHINE WITH INTERTURN
monics generated due to the discrete current distribution in a STATOR FAULT AT SUPPLY DISCONNECTION
squirrel-cage rotor, considering only the fundamental, are given
The modeling technique followed here is similar to [13] and
by
[19]. However, the shorted turns, referred to hereafter as “ ,”
(2) now form another stator circuit whose loop equation and flux
linkage are to be considered. Thus, the loop equations for stator
where is the number of rotor slots. This can be proved by phases including the shorted turns and the rotor can then be
visualizing the rotor bar currents as samples of a continuous si- written as
nusoidal signal with sampling frequency times the funda-
mental space harmonic. The additional frequency components (5)
given by (2) that are generated in the discrete-time signal are
essentially a periodic repetition of the original analog spectrum (6)
[16].
Only some of these harmonics are reflected in the time do- (7)
main as principal slot harmonics (PSH) or rotor slot harmonics
(RSH) whose pole pair numbers match with that of stator space (8)
harmonics [17]. They can be detected as
(9)
(3)
where is the voltage of the machine neutral with respect to
in the line current of a healthy machine [18]. Machine manufac- ground and a null vector of length equal to number of rotor
turing asymmetry, supply voltage unbalance, slotting, and in- bars plus one (due to end ring).
NANDI AND TOLIYAT: TECHNIQUE TO DETECT STATOR INTERTURN FAULTS IN INDUCTION MACHINES 103

The rotor flux linkages , the stator phase flux linkages The stator flux linkages can be derived from (10)–(12) as
, and the stator shorted turns flux linkage are
given by (25)
(26)
(10)
(27)
(11)
(12) Also, from (14) the rotor flux linkage in the absence of the stator
(13) phase currents is
(14)
(28)
where
With , (8), (9), and (19) yield
(15)
(29)
with , or refers to the end ring and
to the rotor loops.
where
Thus, each of the three stator phases and the rotor circuits
will now have one extra term due to the current flowing in these
shorted turns. (30)
The initial values of the faulty and the rotor loop currents can
be found out using the constant flux linkage theorem as follows. with as rotor speed and a rotor position.
Let and be the shorted turns and the rotor flux link- Substituting (25)–(27) in (22)–(24), respectively, and using
ages, respectively, at the instant of supply disconnection. Using (29), can be rewritten as
(13) and (14), they are given by

(16) (31)
(17)
(32)
with
(33)

(18) where
and the stator (including the faulty portion) and rotor
loop currents.
Defining (34)

The expression for the terminal voltages can be obtained from

(19)
the above as

the initial values of the shorted turns and rotor loop cur-
rents, respectively, just after the disconnection, can be computed
as follows: (35)

(20)
(36)
or

(21)
(37)
Since the stator currents are zero after removing the supply, the
phase voltages are given, using (5)–(7), as
IV. SIMULATION RESULTS
(22) The stator interturn fault was simulated using the winding
function approach (WFA) in a manner similar to [11]. Five turns
(23) out of a total of 324 turns (1.5%) in one phase (phase “ ”) were
assumed to be short circuited on a 460-V, three–phase 3–hp
(24) four-pole 36–stator-slot 44–rotor-slot concentrically wound (54
104 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 38, NO. 1, JANUARY/FEBRUARY 2002

Fig. 1. From top to bottom: winding function of healthy phase “a” winding,
faulty and healthy portion of the same winding.

conductors/slot, single layer, coil pitch: 1–12, 2–11, 3–10, 2 Fig. 2. Top: mutual inductance between healthy portion of phase “ a” and rotor
coils in series/phase) star-connected, skewed squirrel-cage in- loop 1; bottom: mutual inductance between faulty portion of phase “a” and rotor
loop 1.
duction machine. The winding function for the healthy portion
of the faulty winding can be expressed as the difference of the
original healthy winding function and the faulty turns’ winding
function. The winding function of the original healthy winding
and the faulty turns can be expressed as a function of stator an-
gular position as

(38)

(39)

respectively. For this particular machine,

and (40)

These winding functions are shown in Fig. 1. In essence,

winding function of a winding for a smooth air-gap machine
represents the magnetomotive force (MMF) distribution in
ampere turns (ATs), produced by that winding along the air gap
with 1 A of current flowing through it. Detailed discussions on
winding function and related topics can be found in [19]. Fig. 2
shows the inductance profile of healthy and the faulty sections
of phase “ ” with rotor loop 1. An exactly similar short circuit
did not produce noticeable performance degradation, increased
noise level, or damage to an identical induction machine on Fig. 3. Simulated, normalized line current spectra of a three–phase 3-hp 60-Hz
skewed 44-bar 460-V four-pole induction motor under healthy (top), voltage
which stator faults experiments were performed repeatedly. unbalance (middle), and stator shorted-turns fault (bottom). The component near
Extensive study on stator faults involving a small number 1205 Hz occurs both under unbalance and stator fault. Unbalance was 2.5% and
of turns [20] indeed reveals that machines with these faults =
5% on two phases. Five out of 324 turns were shorted. Slip 0.042.

are able to run over an extended period without any visible

debilitating effect. Thus, they can be categorized as incipient performed on a healthy, unbalanced % % and a
faults. faulty machine. The plots are shown in Fig. 3. The component
To illustrate that stator faults can be confused with unbal- at 1324.56 Hz is the principle slot harmonic (PSH) related and
anced supply voltage, MCSA on simulated line current was is present in all the cases. The component at 1204.56 Hz occurs
NANDI AND TOLIYAT: TECHNIQUE TO DETECT STATOR INTERTURN FAULTS IN INDUCTION MACHINES 105

Fig. 5. Simulated, normalized line voltage spectra of a three–phase 3-hp 60-Hz

skewed 44-bar 460-V four-pole induction motor with stator shorted-turns fault
Fig. 4. Simulated, normalized line voltage spectra of a three–phase 3-hp 60-Hz at switch-off with different numbers of rotor bars (from top to bottom: R =
skewed 44-bar 460-V four-pole induction motor under healthy (top), voltage 44; 43; 42).
unbalance (middle), and stator shorted-turns fault (bottom) at switch-off.

only under unbalanced and faulty condition. It can be shown

(see the Appendix) that this component is due to the reverse
rotating field. In an actual, healthy machine this component can
also be seen due to inherent constructional imbalance, voltage
unbalance, etc.
The simulated results for terminal voltage spectra for are
shown in Fig. 4. With normalized as 1, the harmonics to
look for can be easily calculated as the 21st or 23rd from (2) or
(4). The 23rd harmonic is present for all three cases because this
is a PSH-related harmonic. However, the 21st harmonic can be
seen only when the stator fault is present.
In order to verify our premise that stator interturn faults in-
deed generate the harmonic numbers given by (2), simulations
were carried out with other numbers of rotor slots. In all cases,
the generated harmonics obeyed the equation; as shown in
Fig. 5.
Fig. 6 shows the spectra of all three line–line voltages of the
machine. The line–line voltage (middle plot), which did not con-
tain the faulty phase, showed the least increase in amplitude for
the 21st harmonic among the three. Thus, the common phase of
the two line–line voltages showing the maximum increase in the
amplitude of the 21st harmonic can be picked out as the faulty
one.
Fig. 6. Simulated, normalized line voltage spectra of a three–phase 3-hp 60-Hz
skewed 44-bar 460-V four-pole induction motor with stator shorted-turns fault
V. EXPERIMENTAL RESULTS at switch-off.
Experiments were performed on a machine identical to the
one with 44 rotor slots. Taps were taken out of the machine was introduced by choosing a lower voltage tap on the trans-
winding in order to perform the tests. The voltage unbalance former supplying the machine. The data were acquired using
106 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 38, NO. 1, JANUARY/FEBRUARY 2002

Fig. 8. Experimental, normalized spectra of line voltages “ab,” “bc,” and “ca”
(top to bottom) of a three–phase 3-hp 60-Hz skewed 44-bar 460-V four-pole
induction motor with stator shorted-turns fault (five turns) at switch-off.

measured harmonic, as can be seen from the FFT under healthy

condition.
Figs. 8 and 9 show the 21st harmonic in the different line–line
voltages with five and six turns shorted. Spectra of line voltage
“ ” (middle plot) that did not include the faulty phase showed
the least increase.
In order to ascertain the consistency of the results, five data
sets of the three line–line voltages were recorded for each of the
healthy, unbalanced, and faulty (one, five, and six turns’ short)
cases. The data were normalized with respect to the fundamental
(harmonic number 1) which was set at 0 dB. The average and
the standard deviation of the five sets for each of these cases are
as shown in Table I (rows 2–6). The line–line voltages that con-
tained the faulty phase “ ” showed a higher 21st harmonic (for
Fig. 7. Experimental, normalized line voltage spectra of a three–phase 3-hp example, 63.85 and 66.32 dB with five turns’ fault) com-
60-Hz skewed 44-bar 460-V four-pole induction motor under healthy (top), ponent, compared to the one ( 71.42 dB) which did not. This
unbalanced (second from the top), and faulty (last three with one, five, and six
turns short, respectively) conditions at switch-off. difference was maximum ( 63.85 71.42 7.57 dB) with
five shorted turns and minimum ( 62.33 63.24 0.91 dB)
with six. This suggests that with a larger number of turns in-
LABVIEW software. Only the first five cycles of data after the volved in faults, even the healthy windings can pick up strong
switch-off were used to compute the spectral components. The signals. However, with six shorted turns, the amplitude of the
sampling frequency was 35 kHz. 21st harmonic component in a fault containing line was max-
Test results obtained under healthy, unbalanced (same as in imum ( 59.53 dB) and least ( 73.57 dB) with one turn shorted.
the simulation) and faulty conditions with one, five, and six Also, the fault possibly cannot be picked up just on the basis
turns shorted are shown in Fig. 7. Although the single turn short of absolute difference in readings between the healthy and the
produced little effect, the 21st harmonic (indicated by the arrow) faulty cases with a small number of turns shorted. In fact, as
shows significant change with five and six turns shorted, clearly mentioned in [20], a small number of shorted turns may actu-
validating the simulation results. The inherent manufacturing ally compensate for the inherent machine asymmetry. This can
asymmetry also did not have any significant effect on the be the cause for reduction in the 21st harmonic in some cases
NANDI AND TOLIYAT: TECHNIQUE TO DETECT STATOR INTERTURN FAULTS IN INDUCTION MACHINES 107

TABLE I
AVERAGE OF THE 21ST HARMONIC MAGNITUDE (NORMALIZED) AND THEIR DIFFERENCES FOR HEALTHY, FAULTY (1, 5, AND 6 SHORTED TURNS) AND
UNBALANCED MACHINE, EACH COMPUTED OVER FIVE DATA SETS FOR THE THREE LINE–LINE VOLTAGES. LINES “ab” AND “ca” CONTAIN THE FAULTY PHASE
(PHASE “a”). THE FIGURES IN BRACKETS GIVE THE STANDARD DEVIATIONS OF THE READINGS.

Fig. 10. Increase in the 21st harmonic component of the line volatge
=
containing the faulty phase with slip 0.0055. Upper: healthy; lower: faulty.

Fig. 9. Experimental, normalized spectra of line voltage “ab,” “bc,” and “ca”
(top to bottom) of a three–phase 3-hp 60-Hz skewed 44-bar 460-V four-pole
induction motor with stator shorted turns fault (six turns) at switch-off.

when a single turn fault was introduced. Moderate deviations

among the 21st harmonic component of the line–line voltages
may, therefore, be an indication of an incipient winding
problem, although these amplitudes themselves may not be
large enough. The standard deviations of all the readings are
quiet low, implying that the results are fairly consistent. The
differences between the readings under the faulty and healthy
conditions (rows 7–9) and faulty and unbalance conditions
(rows 10–12) are also shown. Except for the single turn fault,
voltage unbalance can be clearly distinguished from fault, as
can be seen from readings in rows 11 and 12.
The load dependence of the magnitude of the 21st harmonic
component was also looked into with five shorted turns. Figs. 10 Fig. 11. Increase in the 21st harmonic component of the line voltage with the
and 11 show these harmonics at two widely different values of =
faulty phase. Slip 0.0339. Upper: healthy; lower: faulty.
108 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 38, NO. 1, JANUARY/FEBRUARY 2002

Fig. 12. Decrease in the 21st harmonic component of the line voltage with the
faulty phase as slip increases from near no load to full load.

Fig. 14. Experimental plots showing the increase in 21st harmonic component
in the line (with the faulty phase) voltage at switch-off with 6/324 turns shorted
with the motor operating from a VSI. Top: health; bottom: faulty.

APPENDIX
Fig. 13. Experimental (scaled) line voltage output at switch-off, as obtained The creation of the frequency component at 1204.56 Hz as
from the data acquisition system with the motor operating from a VSI.
shown in Fig. 3 due to reverse rotating magnetic field can be
proved as follows. The proof uses the rotor/stator, stator/rotor
slip at which the machine was operating just before switch-off. coordinate transformation as described in [21]. Stator frame of
Under faulty condition, the 21st harmonic component reduced reference is transformed into rotor frame of reference by adding
only by around 2.2 dB with the higher value of operating slip. rotor speed) to the rotor referenced angular po-
All the earlier experimental results were conducted with the sition. Similarly, rotor frame of reference is transformed into
motor operating at full load (slip 0.042). Fig. 12 shows the stator frame of reference by subtracting from the stator ref-
change in the 21st harmonic with slip. erenced angular position. Only the average part of permeance
To test the suitability of the scheme with voltage-source-in- function (in simple terms, a smooth air-gap machine), , is
verter-fed drives, the motor was also tested with a commer- considered.
cial inverter. The transition of the terminal voltages from the MMFs causing reverse rotating fields (considering only fun-
pulsewidth modulation (PWM) pattern to a sinusoidal one, at damental component of the supply) are of the form
the supply switch-off, can be clearly seen in Fig. 13. Just be-
(A1)
fore switch-off, the inverter fundamental output frequency was
30 Hz. The nature of the terminal voltages after switch-off sug-
where is the number of space harmonic, is the
gests that supply harmonics have little effect on the machine
line frequency in radians per second, and the angular position
terminal voltages once the supply is removed. This is due to the
in the stator frame of reference in mechanical degrees.
fact that the machine terminal voltages after switch-off is pri-
These MMFs acting on produce air-gap flux components
marily dependent on the harmonic content of the magnetizing
that are given by
branch current. Fig. 14 shows the increase of the 21st harmonic
in the line voltage with the fault introduced.
(A2)

with respect to stator, or,

VI. CONCLUSION
A novel stator interturn fault detection scheme by the (A3)
presence of certain harmonics at the terminal voltages of
an induction machine immediately after switch-off was dis- with respect to rotor, being the angular position in the rotor
cussed. The physics behind the production of such harmonics frame of reference.
were explained in detail. Unlike negative-sequence current These components produce additional
or impedance measurements, this technique is insensitive to , pole pair rotor MMF space harmonics [14],
supply voltage unbalance. The faulty phase can be detected, [15], [17] which acting upon produce
too. The very nature of the test also suggests that the supply air-gap flux components of the type
harmonics have little influence on detectability. Corroborative
simulation and experimental results have been presented both (A4)
for line frequency and adjustable-speed drives. (A5)
NANDI AND TOLIYAT: TECHNIQUE TO DETECT STATOR INTERTURN FAULTS IN INDUCTION MACHINES 109

with respect to rotor, or, [11] G. Joksimovic and J. Penman, “The detection of interturn short circuits
in the stator windings of operating motors,” in Proc. IEEE IECON’98,
Aachen, Germany, 31 Aug.–4 Sept. 1998, pp. 1974–1979.
[12] A. Stavrou, H. Sedding, and J. Penman, “Current monitoring for de-
(A6) tecting inter-turn short circuits in induction motors,” in Proc. IEMD’99,
Seattle, WA, May 9–12, 1999, pp. 345–347.
[13] J. Milimonfared, H. M. Kelk, S. Nandi, A. Der Minassians, and H. A.
Toliyat, “A novel approach for broken rotor bar detection in cage in-
(A7) duction motors,” IEEE Trans. Ind. Applicat., vol. 35, pp. 1000–1006,
Sept./Oct. 1999.
with respect to stator. [14] P. L. Alger, The Nature of Induction Machines. New York: Gordon and
Substituting by in (A6) and (A7) leads to Breach, 1965.
[15] M. M. Liwschitz, “Field harmonics in induction motors,” Trans. AIEE,
vol. 61, pp. 797–803, Nov. 1942.
[16] G. Proakis and D. G. Manolakis, Digital Signal Processing, Principles,
Algorithms and Applications, 3rd ed. Englewood Cliffs, NJ: Prentice-
Hall, 1996, p. 271.
(A8) [17] G. Kron, Equivalent Circuits of Electric Machinery. New York: Wiley,
1951.
[18] A. Ferrah, K. J. Bradley, P. J. Hogben-Laing, M. S. Woolfson, G. M
and Asher, M. Sumner, J. Celia, and J. Shuli, “A speed identifier for induc-
tion motor drives using real-time adaptive digital filtering,” IEEE Trans.
Ind. Applicat., vol. 34, pp. 156–162, Jan./Feb. 1998.
[19] , “A method for dynamic simulation of air-gap eccentricity in in-
duction machines,” IEEE Trans. Ind. Applicat., vol. 32, pp. 910–918,
July/Aug. 1996.
(A9) [20] J. Kohler, J. Sottile, and F. C. Trutt, “Condition-based maintenance of
electrical machines,” in Conf. Rec. IEEE-IAS Annu. Meeting, vol. 1,
Phoenix, AZ, Oct. 2–7, 1999, pp. 205–211.
Consider, for example, the case . With [21] D. G. Dorrell, W. T. Thomson, and S. Roach, “Analysis of airgap flux,
Hz, as per (A8), the pole pair number associ- current, vibration signals as a function of the combination of static and
ated with the 1204.56-Hz component dynamic airgap eccentricity in 3-phase induction motors,” IEEE Trans.
Ind. Applicat., vol. 33, pp. 24–34, Jan./Feb. 1997.
is 23 , as noted in Fig. 3. Similarly, from (A9), the
pole pair number associated with the 1324.56-Hz component
is 21 . Since only 23 match Subhasis Nandi (S’97–M’00) received the B.E. de-
a stator space harmonic (given by 5, 7, 11, 13, etc. gree from Jadavpur University, Calcutta, India, the
in general) for a symmetrical three-phase winding), only the M.E. degree from Indian Institute of Science, Ban-
galore, India, and the Ph.D. degree from Texas A&M
1204.56-Hz component can be induced by reverse rotating field. University, College Station, in 1985, 1988, and 2000,
By similar arguments, one can prove that the 1324.56-Hz com- respectively, all in electrical engineering.
ponent is produced by the forward rotating field. The proof can He is currently an Assistant Professor in the De-
partment of Electrical and Computer Engineering,
be carried out by replacing the positive sign before with a neg- University of Victoria, Victoria, BC, Canada.
ative one. Between 1988–1996, he was with TVS Electronics
and the Central Power Research Institute, Bangalore,
India, working in the areas of power electronics and drives. His main research
REFERENCES interests are power electronics and drives and analysis and design of electrical
[1] G. B. Kliman, W. J. Premerlani, R. A. Koegl, and D. Hoeweler, “A new machines, with special emphasis on fault diagnosis.
approach to on-line fault detection in ac motors,” in Conf. Rec. IEEE-IAS
Annu. Meeting, San Diego, CA, Oct. 6–10, 1996, pp. 687–693.
[2] P. J. Tavner and J. Penman, Condition Monitoring of Electrical Ma-
chines, Letchworth, U.K.: Research Studies Press, 1987. Hamid A. Toliyat (S’87–M’91–SM’96) received the
[3] G. Stone and J. Kapler, “Stator winding monitoring,” IEEE Ind. Ap- B.S. degree from Sharif University of Technology,
plicat. Mag., vol. 4, pp. 15–20, Sept./Oct. 1998. Tehran, Iran, the M.S. degree from West Virginia
[4] J. Penman, H. G. Sedding, B. A. Lloyd, and W. T. Fink, “Detection and University, Morgantown, and the Ph.D. degree
location of interturn short circuits in the stator windings of operating mo- from the University of Wisconsin, Madison, in
tors,” IEEE Trans. Energy Conversion, vol. 9, pp. 652–658, Dec. 1994. 1982, 1986, and 1991, respectively, all in electrical
[5] H. A. Toliyat and T. A. Lipo, “Transient analysis of cage induction ma- engineering.
chines under stator, rotor bar and end ring faults,” IEEE Trans. Energy He is currently an Associate Professor in the
Conversion, vol. 10, pp. 241–247, June 1995. Department of Electrical Engineering, Texas A&M
[6] S. Williamson and P. Mirzoian, “Analysis of cage induction motor with University, College Station. His main research
stator winding faults,” IEEE Trans. Power App. Syst., vol. 104, pp. interests and experience include multiphase vari-
1838–1842, July 1985. able-speed drives for traction and propulsion applications, fault diagnosis of
[7] L. Collamati, F. Filippetti, G. Franceschini, S. Pirani, and C. Tassoni, electric machinery, analysis and design of electrical machines, and sensorless
“Induction motor stator fault on-line diagnosis based on labview envi- variable-speed drives. He has authored more than 140 published technical
ronment,” in Proc. MELECON’96, vol. 1, Bari, Italy, May 13–16, 1996,
papers in these fields. He is a member of the Editorial Board of the Electric
pp. 495–498.
Machines and Power Systems Journal.
[8] J. L. Kohler, J. Sottile, and F. C. Trutt, “Alternatives for assessing the
electrical integrity of induction motors,” IEEE Trans. Ind. Applicat., vol. Dr. Toliyat is a Member of the Editorial Board of the IEEE TRANSACTIONS
ON ENERGY CONVERSION, an Associate Editor of the IEEE TRANSACTIONS ON
28, pp. 1109–1117, Sept./Oct. 1992.
[9] C. J. Dister and R. Schiferl, “Using temperature, voltage, and/or speed POWER ELECTRONICS, He is the Vice Chairman of the Electric Machines Com-
measurements to improve trending of induction motor rms currents mittee of the IEEE Industry Applications Society and also serves on several
in process control and diagnostics,” in Conf. Rec. IEEE-IAS Annu. IEEE committees and subcommittees. He is the recipient of the 1996 IEEE
Meeting, vol. 1, St. Louis, MO, 1994, pp. 312–318. Power Engineering Society Prize Paper Award. He received the Texas A&M
[10] M. E. H. Benbouzid, M. Vieira, and C. Theys, “Induction motors’ Select Young Investigator Award in 1999 and the Eugene Webb Faculty Fellow
faults detection and localization using stator current advanced signal Award in 2000. He also received the Space Act Award from NASA in 1999
processing techniques,” IEEE Trans. Power Electron., vol. 14, pp. and the Schlumberger Foundation Technical Award in 2000 and 2001. He is a
14–22, Jan. 1999. member of Sigma Xi.