ACTIVE STATOR WINDING THERMAL PROTECTION FOR AC MOTORS

Pinjia Zhang

Bin Lu

Thomas G. Habetler

Student Member, IEEE

Georgia Institute of Technology
777 Atlantic Dr
Atlanta, GA 30332
USA

Senior Member, IEEE

Eaton Corporation
4201 North 27th Street
Milwaukee, WI 53216
USA

Fellow, IEEE
Georgia Institute of Technology
777 Atlantic Dr
Atlanta, GA 30332
USA

Abstract AC motors are the main workhorses in

process industries. Their malfunction may not only lead
to repair or replacement of the motor, but also cause
significant financial losses due to unexpected process
downtime. Reliable thermal protection of ac motors is
crucial for reducing the motor failure rate and prolonging
a motors lifetime. In this paper, conventional thermal
protection devices and state-of-the-art thermal protection
techniques are reviewed with their advantages and
limitations discussed. Since 2001, the research team at
the authors company and at the Georgia Institute of
Technology has been cooperatively developing a series of
active motor thermal protection methods, as a low-cost
alternative approach to the traditional passive and sensorbased methods. These active thermal protection methods
monitor the average stator temperature via stator
resistance estimation based on the dc equivalent model of
ac motors, using only motor stator voltage and current
measurements. In this paper, an overview of the active
thermal protection techniques for line-started, soft-starterconnected and inverter-fed ac motors is presented, with
their advantages and drawbacks discussed. The active
thermal protection techniques are capable of providing
accurate, and non-invasive thermal protection of ac
motors.
Index Terms AC motor, inverter, starter, stator
resistance, stator temperature, temperature estimation,
thermal protection.

I.

INTRODUCTION

In North America, ac motor systems play a critical role in

industrial process, while consuming more than a half of all the
electric power produced.
Due to their effectiveness,
ruggedness and low maintenance requirements, these ac
motors are widely used in applications of pumps, fans, mills,
compressors, etc. The malfunction of a motor may not only
lead to the repair or replacement of the individual motor, but
also cause significant financial losses due to unexpected
process downtime. Therefore, reliable motor protection is
essential for minimizing the motor failure rate, increasing the
mean time to a destructive motor breakdown, and prolonging a
motors lifetime. Over the years, substantial efforts have been
made in developing preventive monitoring and protection
techniques for ac motors. As an important feature of any
motor protection system, thermal protection is crucial for
avoiding thermal overload and prolonging a motors lifetime.
As one of the major underlying root causes of motor
failures, thermal overload can lead to deterioration of key
components of a motor, including stator winding insulation,
bearing, motor conductors, and core, etc [1-5]. The stator

insulation is typically the most vulnerable component during

thermal overload, since its thermal limit is reached before that
of any other motor component. About 35-40% of induction
motor failures are related to stator winding insulation failure [13]. Stator insulation failures are normally the results of longterm thermal aging [6]. A high stator winding temperature,
which also depends on the insulation class, gradually and
irreversibly reduces the electrical and mechanical performance
of the insulation materials due to chemical reactions, and
eventually leads to insulation failures. The typical thermal
limits of the stator winding for different insulation classes are
listed in Table I [7]. As a rule of thumb, it is estimated that a
motors life is reduced by 50% for every 10C increase above
the stator winding temperature limit. Therefore, a motor must
be de-energized immediately when the thermal limit is
reached. As a result, accurate and reliable thermal protection
of ac motors is essential for prolonging a motors lifetime and
preventing unexpected process downtime.
TABLE I.
TEMPERATURE LIMITS FOR DIFFERENT INSULATION
CLASSES.
Ambient
Rated
Hot Spot
Insulation
Temperature
Temperature
Temperature
Class
(C)
Rise (C)
(C)
A
40
60
105
B
40
80
130
F
40
105
155
H
40
125
180
Motor thermal overload is typically caused by the following
reasons [8]:
1. transient/starting thermal overload;
2. motor overload;
3. unbalanced supply voltages;
4. high ambient temperature;
5. impaired cooling capability.

(a) unbalanced supply voltage

(b) overload

Fig. 1. stator winding damage due to thermal overload [9].

PRESENTED AT THE 2009 IEEE IAS PULP & PAPER INDUSTRY CONFERENCE IN BIRMINGHAM, AL: IEEE 2009 - PERSONAL USE OF THIS MATERIAL IS PERMITTED.

The stator winding damages caused by unbalanced supply

voltage and overload are shown in Fig. 1 (a) and (b),
respectively. It can be observed that the stator winding may
be totally damaged due to thermal overload.
Since the start of the utilization of ac motors, thermal
protection has drawn special attention due to the relatively
severe consequences of stator insulation failures.
To
preventively protect the motor from stator winding insulation
failure and extend a motors lifetime, the stator winding
temperature often needs to be continuously monitored during
operation. Such monitoring is particularly important for
medium- to large-size motors due to relatively larger capital
costs the costs of the associated process downtime due to
unexpected failures. Direct measurements of the stator
winding temperature using embedded thermal sensors, such
as thermocouples, resistance thermal detectors (RTDs),
infrared thermal sensors, etc, are the most reliable approach
for thermal protection. However, their applications are limited
due to economic reasons, especially for small- to medium-size
motors, since the installations of the embedded thermal
sensors are difficult and costly.
Many different types of thermal protection devices have
become commercially available, which by design are not
directly measuring the motor temperature. Their operating
principles and drawbacks are discussed in Section II and III.
The basic concept of an alternative type of thermal protection
techniques: active thermal protection techniques, is presented
in Section IV. The active thermal protection techniques for
line-started, soft-starter-connected, and inverter-fed ac motors,
which are previously proposed by the authors, are summarized
and compared in Sections V, VI, and VIII, respectively. Their
advantages and drawbacks are also discussed. The active
thermal protection techniques are capable of providing
accurate and non-invasive thermal protection of ac motors.

II.

CONVENTIONAL THERMAL PROTECTION DEVICES

Fig. 2. Typical thermal limit curve [10].

The most commonly used conventional motor thermal

protection devices are: dual-element time-delay fuses and
eutectic alloy or bi-metal type overload relays. Motor thermal
protection in industry typically relies on using a combination of
these devices. They are designed based on the thermal limit
curves, which define the safe operating time for different
magnitudes of input currents under both transient and running
overload conditions. The typical thermal limit curve of an ac
motor is shown in Fig. 2 [10].
Although these low-cost thermal protection devices have
been proven to work over the years, they can only roughly
emulate the thermal limit curve.
Besides, even the thermal
limit curve is also only a rough and conservative
representation of the thermal characteristic of ac motors, since
the effects of ambient temperature, unbalanced supply voltage
and the changes of the cooling capability of ac motors are
neglected. In addition, since only the magnitude of the input
current is monitored, these devices are not capable of
protecting motors when high negative-sequence currents are
present. Therefore, nuisance tripping and under-protection are
both common when these devices are used [8, 11].

III.

MICROPROCESSOR-BASED THERMAL OVERLOAD

RELAYS

To overcome the aforementioned drawbacks of

conventional fuses and relays, microprocessor-based thermal
overload protective relays are often applied. They represent
the state of the art in the thermal protection of ac motors.
These microprocessor-based relays typically estimate the
power losses in ac motors using current measurements, and
then calculate the stator winding temperature based on the
thermal model of ac motors. The motor is tripped immediately
when a predetermined maximum permissible temperature is
reached.

Fig. 3. First-order thermal model of ac motors.

The first-order thermal models are the most commonly

used thermal model in this type of relays. An example of the
first-order thermal model is shown in Fig. 3. The stator
winding temperature and the ambient temperature are denoted
as Ts and TA, respectively; Cth represents the equivalent
thermal capacitance, which models the intrinsic thermal
characteristic of the stator winding; Rth represents the
equivalent thermal resistance, which models the cooling
capability from the stator winding to the ambient; Ploss
represents the heat dissipation in the stator winding, which is
mainly the copper loss. Since identification of the thermal
parameters are typically difficult without embedded thermal
sensors, methods have been developed to roughly estimate
the thermal parameters using the motors nameplate
information, including Service Factor and Trip Class [12]. The

first-order thermal model of induction motors is broadly used

due to its implementation simplicity: only the motor current
needs to be measured. However, as these approaches
neglect the temperature rise caused by the other motor losses
and the change of the motors thermal behavior under different
operating conditions, these thermal models are generally
inaccurate and conservative, and therefore often times use of
this approach will trip the motor long before the actual thermal
limit is reached.
To improve the stator temperature estimation accuracy by
considering the effect of different losses on stator temperature,
high-order thermal models are also developed [13-19]. These
methods can model the thermal effects of the losses in
different motor components, such as winding, core, etc, and
therefore are capable of providing more accurate stator
temperature estimation than the first-order thermal models. An
example of a second-order thermal model is shown in Fig. 4
[16], where the stator and the rotor are modeled separately, so
that the thermal effects of the rotor losses and the rotor
temperature rise on the stator temperature rise can be
estimated. The complexity of the thermal model is a tradeoff
between the accuracy of temperature estimation and the
complexity in parameter identification and required
computational effort.
The losses in different motor
components are either calculated online using the current
measurement, or estimated via offline testing. The thermal
parameters in these high-order thermal models are normally
determined based on the measured temperature at different
locations in the motor using embedded thermal sensors. The
applications of high-order thermal models are limited due to
the difficulty in thermal parameter identification, when
embedded thermal sensors are not available.
Another
inherent drawback of the thermal model-based approaches is
their negligence of the changes in the cooling capability of ac
motors. Therefore, using a thermal model with fixed thermal
parameters can lead to rapid motor failures when the cooling
capability of the ac motor is significantly reduced.

motors [20-22]. However, these techniques are mainly used

for improving the field-oriented control performance at low
speed or improving the sensorless speed estimation accuracy.
It is shown in [23] that these approaches are inherently too
sensitive to the variations of motor parameters due to different
operating conditions, and therefore, are not capable of
providing accurate stator resistance estimation for thermal
protection purposes.

IV.

The development of active thermal protection techniques

can be traced back to the early 80s. It is first proposed in [24]
to use the dc model of ac motors to estimate the stator
resistance and temperature, by creating a dc bias in the
motors input voltage and current. Since the dc component
does not pass through the air-gap, stator resistance can be
calculated using only the dc component in the input voltage
and current, as,

Rs =

Fig. 4. Second-order thermal model of ac motors [16].

( R s Rs 0 )
,
T s = T s0 +
Rs 0

(1)

where Ts0 and Rs0 represents the stator temperature and

resistance at room temperature; T s and R s are the estimated

Ts and Rs; and is the temperature coefficient of resistivity.
Many stator resistance estimation approaches have been
proposed based on the equivalent electrical model of induction

dc
2 vab
,
3 iadc

(2)

v dc
i dc
where, ab and a are the dc components of the motor line
voltage vab and phase current ia , respectively, when a dc bias
is injected between phase a and b, phase b and c, of a Yconnected ac motor. However, since power diodes are
typically used to create the dc bias in this approach, dc bias
can only be injected continuously and the magnitude of the
injected dc bias can not be adjusted online, which induces
continuous and uncontrollable torque pulsation and extra heat
dissipation in both the power diodes and also the motor itself.
One novel approach to resolve these design limitations is
by use of a newly developed signal-injection-based active
thermal protection technique. To minimize the torque pulsation
and heat dissipation caused by the injected dc bias,
controllable dc signals are intermittently injected to estimate
the stator temperature. The period of stator temperature
update depends on the requirement of practical applications.
Active thermal protection approaches have been proposed by
the authors for different applications: ac line-started motors
[25, 26], soft-starter-connected motors [27, 28], and inverterfed motors [29]. These techniques are capable of providing
accurate and non-invasive thermal protection of ac motors,
with adaptation to the changes in a motors cooling capability.

V.

Aside from the aforementioned techniques, the stator

temperature can also be monitored based on the estimation of
the stator resistance, since resistance variation is linearly
proportional to temperature variation, as

OVERVIEW OF ACTIVE THERMAL PROTECTION

TECHNIQUES

A.

ACTIVE THERMAL PROTECTION FOR AC LINESTARTED MACHINES

DC Signal Injection for AC Line-started Motors

To inject a controllable dc bias between two phases of ac

motors, a dc injection circuit is proposed in [26]. It consists of
a controllable switch (e.g., a mechanical contactor or a solidstate power switch) and an external resistor (Rext) connected in
parallel, as shown in Fig. 5-(a). In [26], the controllable switch
is implemented using an n-channel enhancement type power
MOSFET. Under dc signal injection mode, the controllable
switch is turned open when ias>0, and turned closed when
ias<0; the equivalent circuits are shown in Fig. 5-(b) and (c).
The asymmetrical resistance causes the voltage drop across
the circuit to be asymmetrical, resulting in injection of a dc
current component into the motor. The magnitude of the

conditions: no load 100%50%75% of the rated load.

The estimated stator resistance with different values of Rext is
shown in Fig. 7-(a), (b) and (c); the measured stator
temperatures with thermocouples at different locations are
shown in Fig. 7-(d).
0.6

0.58

0.58

Estimate of Rs ()

0.6

Estimate of Rs ()

injected dc signal is dependent on the value of Rext, which can

be adjusted depending on the nominal Rs and the rated ias so
that the torque distortion and power dissipation caused by dc
signal injection is within an acceptable level. Under normal
operation, the controllable switch is kept closed to minimize
the heat dissipation in the circuit. Therefore, the effects of dc
signal injection can be minimized by using periodic operation
of the circuit: dc signals are only periodically injected for a
small amount of time, which is sufficient to obtain an accurate
estimate of the stator temperature. As one of the few
laboratory prototypes built to validate the dc injection
approach, a prototype using a commercial mechanical
contactor as the controllable switch is shown in Fig. 5-(d).

0.56

0.56

0.54

0.54

0.52

0.52

No load

0.5

0.5

0.48

Full load Half load 75% load

0

45

90

135

180

225

270

315

0.48

360

No load
0

45

Full load Half load 75% load

90

135

180

225

270

Time (min)

Time (min)

(a) Rext=0.1

(b) Rext=0.2

315

360

90

Estimate of Rs ()

0.58

0.56

0.54

0.52

0.5

0.48

No load
0

45

Full load Half load 75% load

90

135

180

225

270

315

360

Measured stator temperature ( o C)

0.6

T1

80

70

60

50

T2

T3

40

30

20

No load
0

45

Full load Half load 75% load

90

Time (min)

(c) Rext=0.3

135

180

225

270

315

360

Time (min)

(d) measured Ts

Fig. 7. Experimental results (Motor: Leeson, 5 hp, 230 V, 12.4 A,

NEMA-B, ODP, 1745 rpm.)
(d)
Fig. 5. DC injection circuit.

B.

Experimental Testing
Thermocouples

c
b
a

VI.

+ vsw -

DC
Generator
Resistor
Bank

ias
Induction
Motor

230V

A.

MOSFET
Driver

Isolation
Amplifiers

DSP
Board

A/D
Conversion

Load
Field
Voltage

PC
Development
Software

Thermometer

Ts
Rs

Fig. 6. Experimental Setup [26].

A series of lab testing has been conducted to validate this

stator temperature estimation approach using both mechanical
and solid-state switches.
Fig. 6 shows the overall
experimental setup using a MOSFET-based dc injection circuit
[26]. A dc generator supplying a resistor bank serves as the
dynamometer.
Thermocouples are installed at different
locations of the induction motor to measure the stator winding
temperature for comparison purposes. The estimated stator
resistance and the measurement stator temperature are
shown in Fig. 7. The motor is operated under load variation

ACTIVE THERMAL PROTECTION FOR SOFT-STARTERCONNECTED MACHINES

DC Signal Injection using Soft-starters

The soft-starter normally contains multiple anti-parallel

solid-state power switches (e.g., thyristors) to control the
current flow and, in turn, the terminal voltages of the motor.
The soft-starter limits the transient voltages and currents,
reduces the inrush currents, and results in a soft motor start.
After starting, the soft-starter typically enters the bypass
mode, when integrated contactors are closed to minimize the
power dissipation.
A new thyristor gate drive control scheme has been
developed by the authors to inject dc components in the motor
line voltages and phase currents [28].
This scheme
periodically switches the soft-starter operation between two
states after the soft-starter completes the ramp-up mode: a
normal bypass operation state and a dc signal injection state.
During the dc signal injection state, only one contactor
(corresponding to only one phase, e.g., phase a) in the softstarter is kept open, while the other two contactors still work
normally as in bypass mode. Instead of using symmetrical
gate drive signals for all three phases, a short delay is
introduced to the gate drive signal of only the forwardconducting thyristor of phase a (VG1), after the phase a
currents rising zero-crossing. After the dc signal injection
state, the phase a contactor is closed, so that the soft-starter

returns to normal bypass operation state. The equivalent

model of the motor system during dc signal injection state is
shown in Fig. 8. Fig. 9 shows the typical waveforms of the
motor line voltage, vab, phase current, ia, while a small delay
angle of ( < 40) is added. The torque pulsation caused by
the injected dc signal can be controlled under acceptable level
by adjusting the delay angle.

Stator Winding Temperature /C

80

VG1
A
VG2

ia,dc
vab,dc

RS

RS
B

measured Ts
estimated Ts

70

60

50

40

30

RS

20

50

100

150

200

250

300

(a) Normal cooling condition

Fig. 8. DC equivalent circuit of motor, source and soft-starter.
140

(A)

50
0

(V)

0.87

0.88

0.89

0.9

0.91

Stator Winding Temperature /C

-50
0.86

0.92

500
0
-500
0.86

0.87

0.88

0.89

0.9

0.91

0.92

100

80

60

40

Fig. 9. Motor line voltage, phase current during dc signal injection.

B.

measured Ts
estimated Ts

120

Experimental Testing

20

50

100

150

200

250

(b) Cooling fan removed

Fig. 11. Experimental results (Motor: Emerson, 7.5 hp, 230 V, 18.4 A,
NEMA-B, TEFC, 3515 rpm.)

Fig. 10. Experimental setup [27].

The lab setup for the validation of this stator temperature

estimation technique is shown in Fig. 10. The motor is
operated
under
variable
load
conditions:
no
load50%75%100% of the rated load. To test the
feasibility of this approach when the cooling capability of the
motor is reduced, the motors cooling fan is removed under the
same variable load condition.
The estimated stator
temperature and the mean measured temperature using
thermocouples under normal condition and cooling capability
deterioration condition are shown in Fig. 11-(a) and (b),
respectively.

VII.
A.

ACTIVE THERMAL PROTECTION FOR INVERTER-FED

MACHINES

DC Signal Injection using Motor Drives

Open-loop ac motor variable frequency drives are widely

used for controlling the rotor speed of ac motors. In such
drives, motor input voltage commands are generated at a
certain fixed ratio to the frequency (or speed) command to
control the switching of solid-state power switches, e.g. IGBTs.
To inject dc signals, a dc voltage offset can be simply added to
the original input voltage command, which then adjusts the
switching of the power switches automatically. An example of

dc signal injection method using space vector pulse width

modulation (SVPWM) is shown in Fig. 12 [29]. Similar to the
soft-starter dc injection method, this scheme also periodically
switches the drive operation between two states: a normal
operation state and a dc signal injection state. The typical input
current waveform during dc signal injection state is shown in
Fig. 13, where a clear dc component is shown in the motor
current. Since the dc voltage command is constant, the stator
temperature can be estimated using only current
measurements, as,

T s = Ts 0
where

iadc0

1 iadc0 ,
g
iadc

(3)

represents the dc component in the phase current

when the stator temperature is Ts0. Therefore, the voltage

measurement can be avoided, which is highly desirable since
voltage sensors are typically not present in variable frequency
motor drives.

Fig. 14. Experimental setup [29].

Fig. 12. Modified space vector PWM for dc signal injection.

Phase current during dc signal injection
30

with dc signal injection

without dc signal injection

25
phase a current Ia (Amp)

20
15
10
5
0
-5
-10
-15
-20

Fig. 15. Impaired cooling by blocking ventilation.

0

0.02

0.04
0.06
time (sec)

0.08

Fig. 13. Stator current with dc signal injection.

0.1

VIII. SUMMARY OF ACTIVE THERMAL PROTECTION

TECHNIQUES

Stator Temperature Estimation

80
70

Stator Temperature (oC)

60
50
40
30
20
10
0

Measured Ts
Estimated Ts
0

20

40

60
80
time (min)

100

120

140

(a) Normal condition.

Stator Temperature Estimation
80
70

Stator Temperature (oC)

60
50
40
30

IX.

20
10
0

Measured Ts
Estimated Ts
0

50

100

150

time (min)

(b) Blocked ventilation.

Fig. 16. Experimental results (Motor: Leeson, 7.5 hp, 230 V, 20 A,
NEMA-B, ODP, 1760 rpm.)

B.

Conventional passive thermal protection devices,

including dual-element time-delay fuses, eutectic alloy or bimetal type overload relays and microprocessor-based
overload relays, are typically inaccurate due to the difficulty in
the accurate modeling of a motors thermal behavior. The
error in the stator winding temperature estimation may be
more than 30C [28]. Therefore, nuisance tripping is common
when these devices are used.
On the other hand, the active thermal protection
techniques are capable of providing accurate monitoring of the
stator temperature without embedded thermal sensors, as
shown in Section V, VI and VII. The errors in the stator
winding temperature estimation can be smaller than 3C,
depending on the magnitude of the injected dc signals. Only
current and voltage measurements are required, which
guarantees their non-invasive and cost-efficient nature. The
torque pulsation and extra heat dissipation caused by dc signal
injection can be controlled within acceptable levels. The dc
signals are intermittently injected to minimize the effects of the
dc signal injection. The period of dc signal injection depends
on the frequency of stator temperature estimation updates
required by the practical application, given a motors typical
thermal time constant.
Because of their accuracy in the stator winding
temperature estimation, their robustness to the variations in
the electrical/thermal parameters of ac motors and their
capability of adaptation to the changes in a motors cooling
capability, these active thermal protection techniques are
highly preferred over conventional thermal protection
techniques for the reliable thermal protection of ac motors.

Experimental Testing

The lab setup for the validation of this approach is shown

in Fig. 14. The motor is then operated under variable load
conditions: no load100%50%75% of the rated load. To
test the effectiveness of this approach when the motors
cooling capability is deteriorated, the ventilation is blocked,
when the motors is operated under the same variable load
condition, as shown in Fig. 15.
The estimated stator
temperature and the average temperature measured using
thermocouples under normal condition and cooling capability
deterioration condition are shown in Fig. 16-(a) and (b),
respectively. The value of the stator resistance estimated
using this technique can also be further used for improving the
field-oriented control performance at low speed or improving
the sensorless speed estimation accuracy.

CONCLUSIONS

Thermal protection of ac motors is crucial for preventing

catastrophic motor breakdown, prolonging motors lifetime, and
avoiding the extraordinary financial losses due to unexpected
industrial process downtime caused by motor failures. A
review of available thermal protection devices and techniques
has been presented in this paper. The principle and drawback
of conventional overload relays and microprocessor-based
thermal relays have been discussed. It has been concluded
that these thermal protection techniques are not capable of
providing reliable thermal protection for ac motors under
different operating conditions, due to:
1. inaccurate modeling of the thermal characteristics of
ac motors under different operating conditions;
2. negligence of the changes in a motors cooling
capability.
To overcome the drawbacks of conventional thermal
protection techniques, an alternative type of thermal protection
technique: active thermal protection, has been proposed by
the authors for ac line-started, soft-starter-connected, and
inverter-fed ac motors. By intermittently injecting dc signals
into the motors, the dc model of ac motors has been used for
the estimation of the stator winding resistance and
temperature. It has been clearly shown that these techniques
are capable of providing accurate monitoring of the stator
temperature without embedded thermal sensors. The errors in
the stator winding temperature estimation can be within 3C,
depending on the magnitude of the injected dc signals. The
torque pulsation caused by the injected dc signals can be

controlled under acceptable level. The importance of these

techniques lies in their non-invasive nature:
1. only current and voltage measurements are required;
2. motors normal operation is not disturbed.
Therefore, the active thermal protection techniques are
capable of providing non-intrusive, low-cost and reliable
thermal protection of ac motors, when moderate torque
pulsation is acceptable for the practical applications.

X.

ACKNOWLEDGEMENT

The authors would like to thank Dr. Sang-Bin Lee for his
early contribution in this research project. This work was
financially supported in part by the U. S. Department of Energy
under Grant DE-FC36-04GO14000 and such support does not
constitute an endorsement by DOE of the views expressed in
the article.

XI.
[1]

[2]

[3]

[4]

[5]

[6]

[7]

[8]
[9]
[10]
[11]

REFERENCES

"Report of Large Motor Reliability Survey of Industrial

and Commercial Installations, Part II," Industry
Applications, IEEE Transactions on, vol. IA-21, pp.
865-872, 1985.
"Report of Large Motor Reliability Survey of Industrial
and Commercial Installations, Part I," Industry
Applications, IEEE Transactions on, vol. IA-21, pp.
853-864, 1985.
"Report of Large Motor Reliability Survey of Industrial
and Commercial Installations: Part 3," Industry
Applications, IEEE Transactions on, vol. IA-23, pp.
153-158, 1987.
P. F. Albrecht, J. C. Appiarius, R. M. McCoy, E. L.
Owen, and D. K. Sharma, "Assessment of the
reliability of motors in utility applications- Updated,,"
Energy Conversion, IEEE Transaction on, vol. EC-1,
pp. 39-46, 1986.
A. H. Bonnett and G. C. Soukup, "Cause and analysis
of stator and rotor failures in three-phase squirrelcage induction motors," Industry Applications, IEEE
Transactions on, vol. 28, pp. 921-937, 1992.
G. C. Stone, I. M. Culbert, and B. A. Lloyd, "Stator
insulation problems associated with low voltage and
medium voltage PWM drives," in Cement Industry
Technical Conference Record, 2007. IEEE, 2007, pp.
187-192.
"Information Guide for General Purpose Industrial AC
Small and Medium Squirrel-Cage Induction Motor
Standards," NEMA Standard MG1-2003, August
2003.
S. F. Farag, R. G. Bartheld, and T. G. Habetler, "An
integrated on-line motor protection system," Industry
Applications Magazine, IEEE, vol. 2, pp. 21-26, 1996.
"Failures in three-phase stator windings," West Coast
Rewind and Electrical Control Pty Ltd, available at
http://www.westcoastrewind.com.au/repair.php.
"IEEE guide for the presentation of thermal limit
curves for squirrel cage induction machines," IEEE
Standard 620-1996, August 1996.
R. J. Brighton and P. N. Ranade, "Why Overload
Relays Do NOT Always Protect Motors," Industry
Applications, IEEE Transactions on, vol. IA-18, pp.
691-697, 1982.

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

M. S. Abou-El-Ela, A. I. Megahed, and O. P. Malik,

"Thermal model based digital relaying algorithm for
induction motor protection," in Electrical and
Computer Engineering, 1996. Canadian Conference
on, 1996, pp. 1016-1019 vol.2.
A. Boglietti, A. Cavagnino, M. Lazzari, and A.
Pastorelli, "A simplified thermal model for variable
speed self cooled industrial induction motor," in proc.
37th IAS Annual Meeting, 2002. , 2002, pp. 723-730
vol.2.
A. Bousbaine, M. McCormick, and W. F. Low, "In-situ
determination of thermal coefficients for electrical
machines," Energy Conversion, IEEE Transaction on,
vol. 10, pp. 385-391, 1995.
D. R. G. Champenois and D. S. Zhu, "Electrical and
thermal performance predictions in inverter-fed
squirrel-cage induction motor dirves," Electric
Machines and Power Systems, vol. 22, pp. 355-369,
May/June 1994.
Z. Gao, T. G. Habetler, R. G. Harley, and R. S. Colby,
"A Sensorless Adaptive Stator Winding Temperature
Estimator for Mains-Fed Induction Machines With
Continuous-Operation Periodic Duty Cycles," Industry
Applications, IEEE Transactions on, vol. 44, pp.
1533-1542, 2008.
K. D. Hurst and T. G. Habetler, "A thermal monitoring
and parameter tuning scheme for induction
machines," in proc. 32nd IAS Annual Meeting, IAS
'97, 1997, pp. 136-142 vol.1.
J. F. Moreno, F. P. Hidalgo, and M. D. Martinez,
"Realisation of tests to determine the parameters of
the thermal model of an induction machine," Electric
Power Applications, IEE Proceedings -, vol. 148, pp.
393-397, 2001.
P. S. H. Nestler, "On-Line estimation of temperatures
in electrical machines by an observer," Electric
Machines and Power Systems, vol. 21, pp. 39-50,
1993.
T. G. Habetler, F. Profumo, G. Griva, M. Pastorelli,
and A. Bettini, "Stator resistance tuning in a statorflux field-oriented drive using an instantaneous hybrid
flux estimator," Power Electronics, IEEE Transactions
on, vol. 13, pp. 125-133, 1998.
G. Guidi and H. Umida, "A novel stator resistance
estimation method for speed-sensorless induction
motor
drives,"
Industry
Applications,
IEEE
Transactions on, vol. 36, pp. 1619-1627, 2000.
C. B. Jacobina, C. C. de Azevedo, M. N. Lima, and L.
A. de Souza Ribeiro, "Online estimation of the stator
resistance and leakage inductance of a four-phase
induction machine drive," Power Electronics, IEEE
Transactions on, vol. 19, pp. 10-15, 2004.
S.-B. Lee, T. G. Habetler, R. G. Harley, and D. J.
Gritter, "An evaluation of model-based stator
resistance estimation for induction motor stator
winding temperature monitoring," Energy Conversion,
IEEE Transaction on, vol. 17, pp. 7-15, 2002.
D. A. Paice, "Motor Thermal Protection by
Continuous Monitoring of Winding Resistance,"
Industrial Electronics and Control Instrumentation,
IEEE Transactions on, vol. IECI-27, pp. 137-141,
1980.
S.-B. Lee and T. G. Habetler, "A remote and
sensorless thermal protection scheme for small line-

[26]

[27]

[28]

[29]

connected ac machines," Industry Applications, IEEE

Transactions on, vol. 39, pp. 1323-1332, 2003.
S.-B. Lee and T. G. Habetler, "An online stator
winding resistance estimation technique for
temperature monitoring of line-connected induction
machines," Industry Applications, IEEE Transactions
on, vol. 39, pp. 685-694, 2003.
P. Zhang, Y. Du, B. Lu, and T. G. Habetler, "A
Remote and Sensorless Thermal Protection Scheme
for Soft-Starter-Connected Induction Motors," in
Industry Applications Society Annual Meeting, 2008.
IAS '08. IEEE, 2008, pp. 1-7.
P. Zhang, B. Lu, and T. G. Habetler, "A Remote and
Sensorless Stator Winding Resistance Estimation
Method for Thermal Protection of Soft-StarterConnected
Induction
Machines,"
Industrial
Electronics, IEEE Transactions on, vol. 55, pp. 36113618, 2008.
P. Zhang, B. Lu, and T. G. Habetler, "An Active Stator
Temperature Estimation Technique for Thermal
Protection of Inverter-Fed Induction Motors with
Considerations of Impaired Cooling Detection," in
Electric Machines & Drives Conference, IEEE
International, IEMDC '09, 2009.