ACTIVE STATOR WINDING THERMAL PROTECTION FOR AC MOTORS

Pinjia Zhang
Student Member, IEEE Georgia Institute of Technology 777 Atlantic Dr Atlanta, GA 30332 USA

Bin Lu
Senior Member, IEEE Eaton Corporation 4201 North 27th Street Milwaukee, WI 53216 USA

Thomas G. Habetler
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.

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.

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

(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.

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

II.

CONVENTIONAL THERMAL PROTECTION DEVICES

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.

OVERVIEW OF ACTIVE THERMAL PROTECTION TECHNIQUES

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 =

dc 2 vab , dc 3 ia

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

ACTIVE THERMAL PROTECTION FOR AC LINESTARTED MACHINES

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

A.

DC Signal Injection for AC Line-started Motors

( 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

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

Estimate of Rs ()

0.56

0.54

0.52

No load

0.5

Estimate of Rs () Full load Half load 75% load

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).

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 0.6

0.58

0.58

0.56

0.54

0.52

0.5

0.48 0 45

0.48 0

No load
45

Full load Half load 75% load

90 135 180 225 270 315 360

90

135

180

225

270

315

360

Time (min)

Time (min)

(a) Rext=0.1
0.6

(b) Rext=0.2
90

Measured stator temperature ( o C)

0.58

80

T1

Estimate of Rs ()

0.56

70

0.54

60

50

0.52

T2
40

T3

0.5

30

0.48 0

No load
45

Full load Half load 75% load

90 135 180 225 270 315 360

20

No load
0 45

Full load Half load 75% load

90 135 180 225 270 315 360

Time (min)

Time (min)

(c) Rext=0.3

(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.

VI.
DC Generator Resistor Bank

B.
c b a

Experimental Testing
Thermocouples

ACTIVE THERMAL PROTECTION FOR SOFT-STARTERCONNECTED MACHINES

DC Signal Injection using Soft-starters

A.

+ vsw 230V

ias Induction Motor Load Field Voltage

MOSFET Driver PC Development Software DSP Board

Isolation Amplifiers A/D Conversion

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

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

Stator Winding Temperature /C

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.
VG1 A VG2 vab,dc RS B C ia,dc RS RS

80 measured Ts estimated Ts

70

60

50

40

30

20

50

100

150

200

250

300

(a) Normal cooling condition Fig. 8. DC equivalent circuit of motor, source and soft-starter.
50
140 measured Ts estimated Ts

(A)

0 -50 0.86 0.87 0.88

120
Stator Winding Temperature /C

0.89

0.9

0.91

0.92

100

(V)

500 0 -500 0.86 0.87 0.88 0.89 0.9 0.91 0.92

80

60

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

40

B.

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
dc ia 0

dc 1 ia 0 , g dc ia

(3)

represents the dc component in the phase current

Fig. 14. Experimental setup [29].

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

Phase current during dc signal injection 30 25 20 phase a current Ia (Amp) 15 10 5 0 -5 -10 -15 -20 0 0.02 0.04 0.06 time (sec) 0.08 0.1 with dc signal injection without dc signal injection

Fig. 15. Impaired cooling by blocking ventilation.

Fig. 13. Stator current with dc signal injection.

Stator Temperature Estimation 80 70 60 Stator Temperature (oC) 50 40 30 20 10 0 Measured Ts Estimated Ts 0 20 40 60 80 time (min) 100 120 140

VIII. SUMMARY OF ACTIVE THERMAL PROTECTION TECHNIQUES

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.

(a) Normal condition.

Stator Temperature Estimation 80 70 60 Stator Temperature (oC) 50 40 30 20 10 0 Measured Ts Estimated Ts 0 50 time (min) 100 150

IX.

CONCLUSIONS

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

B.

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.

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.

[12]

[13]

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.

[14]

[15]

XI.
[1]

REFERENCES

[2]

[3]

[4]

[5]

[6]

[7]

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

"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.

[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.