Unbalance and Harmonics Detection in Induction ... - IEEE Xplore

IEEE SENSORS JOURNAL, VOL. 6, NO. 3, JUNE 2006

605

Unbalance and Harmonics Detection in Induction Motors Using an Optical Fiber Sensor Jesus M. Corres, Javier Bravo, Francisco J. Arregui, Member, IEEE, and Ignacio R. Matias, Senior Member, IEEE

Abstract—In this work, a new method for the detection of the negative effects of a particular unbalanced voltage and inverter harmonics on the performance of an induction motor using fiber sensors is proposed. Supplying a three-phase induction motor with unbalanced voltages causes an oscillating electromagnetic torque that generates vibrations, increased losses, efficiency reduction, and an extra temperature rise that leads to a reduction on insulation life of the machine. A new in-line fiber etalon accelerometer has been designed to detect these vibrations in the range DC-500 Hz. The in-line fiber etalon scheme used provides high robustness and stability, giving enough sensitivity to monitor the low-frequency and low-amplitude oscillations in the stator of the machine that exist in a voltage unbalance situation. To prove this claim, a 1.5–kW squirrel cage induction motor is analyzed under different unbalance levels. It is shown that a precise unbalance factor can be detected without accessing to the electric part of the machine and an accurate monitoring can be obtained using the high-resolution analysis proposed. Index Terms—Engine fault detection, Fabry-Pérot interferometers, ILFE accelerometer, induction motor, unbalanced voltage.

I. INTRODUCTION CCORDING to the U.S. Department of Energy (DoE), 70% of the electricity is used by industrial motors, and in a typical industry the 80% of the load consists of three-phase AC induction motors [1]. Induction motors are being used more than ever before due to their versatility, dependability and economy. In most cases, condition monitoring schemes have focused on one of three induction motor components: the stator, rotor or bearings. Actually vibration monitoring is accepted for diagnosis of electrical nature faults for many types of machines since this practice does not require either modification of the machine or the access to the supply lines [2]–[4]. Large systems are often equipped with mechanical sensors, primarily vibration sensors based on proximity probes. Those however are delicate and expensive. While the use of vibration monitoring is currently extensive, moderately little attention has been paid to voltage unbalance in the motor supply. Voltage unbalance is related with the case of an unbalanced distribution system network or the use of single phase loads. Some representative examples are unsymmetrical

A

Manuscript received March 14, 2005; revised July 4, 2005. This work was supported by the Spanish Ministerio de Ciencia y Tecnologia-FEDER under Research Grant CICYT-TIC 2003-00909 and Gobierno de Navarra Research Grants. The associate editor coordinating the review of this paper and approving it for publication was Dr. Subhas Mukhopadhyay. The authors are with the Universidad Pública de Navarra. Campus Arrosadía s/n. Edificio Los Tejos. Departamento Ing. Eléctrica y Electrónica, 31006-Pamplona (Navarra) Spain ([email protected]; [email protected]; [email protected]; [email protected]; www.unavarra.es). Digital Object Identifier 10.1109/JSEN.2006.874441

transformer windings or transmission line impedance, unbalanced three-phase loads, blow fuses on three-phase systems, or switching significant single-phase loads, just to mention a few. Because of the wide use of the induction motors both in industrial and residential areas, the damaging effects on induction motors will cause an important economic impact. Under a voltage unbalance the amplitude of the line currents deviation can be many times larger than the voltage change, creating a torque pulsation. From the electrical viewpoint, vibrations are due to varying magnetic forces acting between the stator and rotor of the machine. Mechanical vibration of electric machines is a source of troubles because of the faster ageing of the machine. Another cause of machine’s life reduction is the extra temperature rise that the induction motor under voltage unbalance experiments because of the efficiency reduction since even a small unbalance in the system severely increases the rotor losses. When functioning at low speed the thermal dissipation capability is low, and the derating can become particularly serious. International Electrical Code (IEC) standards [5] restrict the permissible voltage unbalance on induction motors to 1% and obligate to derate if voltage unbalance is greater (5%). The causes for lowering the machine power rating exist in all stages of the electrical energy transformation, from the generation to the final user. As it is well known the reliability of power electronics systems is of paramount importance in industrial, commercial, aerospace and military applications. One of the main problems in pulsewidth modulation (PWM) drives is the current distortion because of the nonideal characteristics of GTOs, IGBTs, or MOSFETs in the inverter bridge. The main source of output current distortion, and consequently in the torque ripple in the motor shaft, is the dead time insertion that must be present in order to prevent the short circuit of the inverter bridge. The use of microprocessors gives high quality control but, although the PWM waveforms are becoming more and more complex, these waveforms rarely have the same shape on each phase and, therefore, the three-phase output voltages can be unbalanced. Also, it is well known that from the safety point of view, preventive maintenance, for example of critical reactor components in nuclear plants, like the refrigeration pumps, has a significative importance. A common characteristic of these large machines, that are the target of this paper, is the presence of high current and voltage levels that create high derivatives and electromagnetic interferences. When switching high power levels for machine control, the commutation device can cause unacceptable interferences in the measured signal. Because of that, the electro-magnetic interference (EMI) immunity of fiber-optic sensors is an advantageous candidate to cover the sensing needs in these fields of application. The high electrical

1530-437X/$20.00 © 2006 IEEE

606


Below, the method of the symmetrical components is exposed. It is used here to calculate the motor torque under different voltage unbalance degrees. A. Symmetrical Components

Fig. 1. Schematic diagram of the electro-mechanical system.

isolation allows the use of fiber-optic probes in places where the use of electrical instrumentation is dangerous or inoperative, such as under highly explosive atmospheres, for example in the petrochemical industry. There are several fiber-optic based techniques for vibration measurement [6]. Widely speaking they can be classified into fiber Bragg gratings (FBGs) [7], intensity modulation [8], and Fabry-Pérot interferometry (FPI) [9]. The highest sensitivity is provided by the interferometric techniques, and among them the most attractive probably is the FPI, because of its easy implementation, low size and robustness. The intrinsic configuration [10] uses semi reflective surfaces on the faces of a segment of fiber fused to the lead fibers in order to create the interferometric cavity. The extrinsic configuration [11] consists of a micrometric alignment tube that provides mechanical stability to the pair of cleaved fiber end faces. The main disadvantage of this scheme comes from the difficulty to protect adequately the cavity in order to obtain robust and long term usable sensors. An alternative configuration that produces robust devices, used in this study, is the in-line fiber etalon (ILFE) proposed by Sirkis et al. [12], [13]. This sensor is built by fusing a segment of hollow core fiber between two standard single-mode fibers to create the interferometric cavity. As the outer diameter of the hollow core segment match with the single-mode fibers, the stress at the discontinuities is significantly low. The sensor used in this work is based on the seismic mass effect with two different ranges of operation. Below the resonance frequency it works as an accelerometer, and for higher frequencies the output is proportional to the displacement, working as a vibrometer. The paper structure will start with some definitions about the system under voltage unbalance. Then the transducer construction and the main properties of its performance will be shown. After that, the experimental results will be presented and the capacity of the proposed system to detect the electrical unbalance in three-phase induction machines will be tested. II. THEORETICAL BACKGROUND The system under study is schematized in Fig. 1. The principle of operation consists of the presence of an electromagnetic motor torque, created by the angle between the rotor and stator flux angles of any electrical machine, which is present both in the stator and rotor of the machine. The torque on the stator serves to induce the change in the speed of the load, as required by the power controller. Part of this torque can be detected using a gauge torque transducer in the couplings with the load. The reaction torque is transmitted by the mechanical assembly to the supporting base, inducing undesirable oscillations that can be detected with the use of adequate vibrometers.

Classical fault analysis of unbalanced power systems has utilized a symmetrical component based approach [14] to obtain the zero, positive and negative sequence model for system components. Given the sequence line to neutral voltages , the sequence voltages are computed with [15]

(1) . As zero-sequence components are never where present in the line-to-line voltages regardless of the unbalance level, negative sequence voltage is the primary cause of voltage unbalance. The electrical model of [16] can be reduced further to the electrical networks shown in Fig. 2, for the positive and negative sequences. The negative sequence equivalent circuit has input impedance similar to that of the locked rotor, in which even a small amount of negative sequence voltage causes a large current and the stator losses increase. The electromagnetic torque is given by the following expression [17]:

(2) where is the synchronous electrical frequency, is the rotor is the slip. electrical frequency, and B. Voltage Unbalance Definitions Comprehension of how voltage unbalance influences the steady-state performance of induction motors is the theoretical root of standards related to induction motor operation and protection. The voltage unbalance factor (VUF) of a sinusoidal three-phase voltage system is defined by the IEC as the quotient between the negative and positive sequence components [18] (3) Normally positive sequence voltage is very close to the rated voltage, and the negative sequence voltage is very close to the VUF. Thus, VUF can indeed be considered as the negative sequence component per unit. Also, the phase voltage unbalance rate (PVUR) is defined in the IEEE Standard 141

(4) where

.

CORRES et al.: UNBALANCE AND HARMONICS DETECTION IN INDUCTION MOTORS USING AN OPTICAL FIBER SENSOR

607

Fig. 2. Sequence models of the induction motor.

Fig. 3. Schematic diagram of ILFE.

C. Harmonics Influence on the Sensing Scheme Harmonics produced by inverters can be in some cases a main source of vibration and oscillations in AC motors. The fundamental harmonic component is positive-sequence, second harmonic is negative-sequence, and third harmonic is zero-sequence, repeating this pattern higher order harmonics. Negative sequence harmonics produce a pulsating torque that causes excess of power losses in the winding as it works against the fundamental component. This is the case of the harmonic component created by the voltage unbalance. Those with positive sequence have a forward rotation and also increase the losses due to skin effect and eddy currents. Finally, triplen harmonics are zero-sequence harmonics that are multiples of third harmonic and they have no effect on torque pulsation. The harmonics caused by inverters, in practice, are odd and not a multiple of three. In [19] the impact of harmonics on induction motors is analyzed, and for example, for the lowest harmonic frequencies, which are fifth and seventh in a threephase square wave inverter, they produce a torque that pulsates at a sixth harmonic frequency. In addition, as higher order harmonics produce torque components at higher frequencies, they have a lower impact on the speed ripple because of the integrative effect of the system inertia. The harmonic content of the inverter output depends on the type of control. For the six-step drives, which basically have a square wave output, torque ripple is a problem, because at low frequencies the stepped nature of the stator rotating field causes the torque to be applied in pulsations. However, on modern sinusoidal PWM the problem of torque pulsations is greatly reduced, because the output wave is near-sinusoidal, even at very low speeds. In [20] we can see how the PWM modulation pushes the harmonics into a high-freand its sidequency range around the switching frequency bands (5–10 kHz). A PWM drive produces little heating and torque precisely because of the high frequency. The leakage inductance of the motor goes up in proportion to frequency, so higher frequency voltage harmonics result in very little current and very little torque and heating effect.

As far as we know, the second order torque ripple harmonic is not involved in the electromagnetic nature oscillations caused by the switching nature of power inverters, unless when the output have voltage unbalance, for example caused by a deadtime problem. Nevertheless, if a symmetrical system with harmonics becomes unbalanced, a second harmonic negative-sequence current will flow independent of the existing harmonics. The resulting unbalanced (negative sequence) currents induce double system frequency currents in the rotor that quickly cause rotor overheating. Under voltage unbalance, thermal, mechanical and magnetic stresses appear and cause the majority of rotor faults, which finally are due to broken rotor bars. In [21], a simulation study of an induction motor with rotor broken bars is reported. The spectrum of the torque ripple, and in consequence of the vibration , components, where is the speed slip spectrum, gives and is the line frequency. As the slip is low in normal operation, the vibration harmonics are in the low-frequency range, around 4 Hz. III. SENSOR FABRICATION AND CHARACTERIZATION The seismic mass effect is the principle of operation of the developed sensor, where the system is created by the optical fiber and a concentrated mass. The optical fiber works as the spring in the spring-mass system, and its deformation is registered using interferometry of the light reflected by the mirrors of a Fabry-Pérot etalon. The etalon scheme is shown in Fig. 3. The first step for the fabrication of the sensor consists of cleaving a standard singlemode fiber and a 50/125 hollow core fiber. As the cleaving angle is of vital importance in order to maintain the amount of light that reaches the detector, each cleaved fiber is tested launching the beam of a laser source and measuring the power of the reflected light. Then the cleaved end face of the single-mode-fiber (SMF) is fused to the hollow core. The next step consists of cutting the hollow core fiber attached to the SMF, setting at this

608


Fig. 4. Schematic diagram of unbalance monitoring system.

point the length of the cavity. The final step is the fusion of the hollow core fiber end face to another single-mode fiber. This technique has the advantage that both mirrors are protected against external factors and parallelism between both surfaces can be fitted with high accuracy. Nevertheless, due to the high sensitivity required for the measurement of rotating machine vibrations, it has been necessary to build sensors with a cavity on the millimeter range. The schematic setup of the measurement system is shown in Fig. 4. The light from a 1-mW laser at 1310 nm passes through one input of a 2 1 SMF coupler. The output arm is attached to the ILFE sensor and the other input is connected to a photodetector that detects the light reflected by the sensor containing the vibra. The photodetector tion information, with a gain of 10 is connected to an oscilloscope with a FFT module to show the frequency domain representation of the vibration applied signal. cavity length with a lead concentrated mass An ILFE of 500of 40 g bonded to the end of the fiber was used. When acceleration is impressed to the frame where one point of the fiber is bonded, a strain caused by the inertial force appears in all the sections of the fiber. Particularly in the hollow core fiber the strain causes a deformation, changing the length between both ends of the cavity. Also, the deformation generates a change in the optical power reflected because of the light interference. This is basically the operation of the proposed sensor. The ILFE sensor exhibits a good linearity in the measurement range, as it can be seen in Fig. 5. The output can be considered approximately linear in the range 20%–80%, with a typical deviation of less than 2%. The minimum detectable acceleration depends on the sensor range and on the amplifier of the signal conditioning system. The limiting factor is the resolution of the photodetector signal conditioning. The optical-power meter has . The minimum detectable an inherent random noise of 8.8 amplitude acceleration was assumed when the noise was of the same amplitude as the output of the sensor. In this situation the sensor was put under a sinusoidal oscillation using the electrodynamic shaker. However, using an appropriate filter, the noise can be reduced consequently increasing the signal-to-noise ratio (SNR) of the recovered signal. The minimum detectable accel, which corresponds eration amplitude was found to be 0.5 to a 300-nm displacement amplitude, measured at 200 Hz. Hysteresis was negligible.

Fig. 5. Amplitude response of the optical fiber accelerometer at f = 200 Hz with a sensitivity of 1.76 nWs m .

It has to be taken into account that the fiber is an elastic material that can be modeled as a spring, with the recovery constant , given by the area of the hollow core section and the glass Young’s module. Also, the presence of damping is useful to reduce the impact of resonances and stabilize the mass-spring system. Therefore, applying modal analysis to the proposed mechanical system for the steady-state sinusoidal movement, the amplitude of the sinusoidal deformation of the fiber, is given by [22] (5) where , , . The variis able is the frequency of the sinusoidal vibration wave, is the the inertial mass in Kg, is the damping in kg/s, and amplitude of the absolute displacement of the mounting. This , where is displacement amplitude is given by the acceleration amplitude. According to this mathematical model the frequency response of the sensor has been simulated and the results can be seen in Fig. 6, where a resonance peak around 85 Hz can be appreciated. For frequencies below the resonance frequency, the output of the sensor is proportional to the acceleration, even for DC values. The resonance frequency depends on several factors, but the main variable for its adjustment is the overall length of the fiber that is between the concentrated mass and the frame to which it is bonded. The resonance peak establishes the working range as accelerometer and as vibrometer for the


609

Fig. 6. Theoretical analysis of the frequency response of the ILFE sensor. Fig. 7. Motor test-rig with 1.5-kW three-phase induction machine.

sensor. When working as a vibrometer (for frequencies three times higher than the resonance), the relative displacement between the mass and the base (sensed by the transducer) is essentially the same as the displacement of the base. In this case, the output of this sensor is proportional to the absolute displacement of the machine. In order to extend the measurement range of the system, an easy signal processing of the spectrum can be used to obtain the same magnitude (acceleration, velocity or displacement) in all the measuring range, whenever the worst-case SNR (for the highest frequency) be adequate for the application of interest.

TABLE I EXPERIMENTAL TEST-RIG PARAMETERS

IV. EXPERIMENTAL RESULTS AND DISCUSSION Based on the theoretical studies explained in [22], and on both theoretical and experimental results, a linear relationship between the current harmonics and vibration level can be assumed. However it is important to note that the vibrations tends to be nonstationary in time and due to the complex relationship between the stator frame and the airgap permeance the amplitudes of the harmonics can behave in a different way. This relationship is highly dependent on the mechanical system, as it differs greatly for each vibration frequency. A robust and well-balanced test-rig, implemented in our laboratory and shown in Fig. 7, was used. Nominal and electrical model parameters of the AC induction motor are shown in Table I. With the use of a TMS320C30 digital signal processor control card and IGBT inverter, any failure of electro-magnetic origin that can exists in a real machine, can be recreated on the test-rig because it controls the instantaneous electromagnetic flux and, in consequence, the instantaneous torque applied. More details about the experimental test can be found in [23]. A. Test of the System With Square Waveform In order to verify the capacity to detect vibrations generated by a failure of the electrical system, a twin machine has been employed to introduce an electromagnetic disturbance torque. An electromagnetic torque square waveform of 3.5 nm was introduced using the DSP-based field oriented vector control algorithm [23]. In spite of the high test-rig robustness, the Fig. 8 shows how the ILFE sensor detected the 10-Hz fundamental and its main harmonics. As expected, odd harmonics are detected due to the square disturbance waveform applied with the load motor. The most

Fig. 8. ILFE sensor output (FFT), 3.5-nm square-wave disturbance torque applied.

important it is the third harmonic because the electromagnetic field created by this component of the flow in the rotor of the induction machine produces as well a torque that it opposes to the direction of rotation of the fundamental component. B. Three-Phase Voltage Unbalance Test For this test, the induction motor has been fed with threephase sinusoidal waveforms. A variable degree of voltage unbalance on phase was introduced reducing the amplitude of the waveform. The state of the machine was logged measuring the shaft speed, the torque given by the gauge transducer and the electrical magnitudes (currents and voltages). The power factor was also monitored. Initially, a balanced three, 45 Hz) was applied. Then the phase voltage system (200

610

Fig. 9. Electrical performance of the induction machine under voltage unbalance (0%–10%).


Fig. 11. Transient response to a voltage unbalance PVUR (0%–5%).

Fig. 10. Supply voltage waveform (200 V , 45 Hz) applied and filtered ILFE vibration waveforms obtained under unbalance operation (5%).

PVUR was increased in 0.5% steps up to a 10%, reducing for this the amplitude of . In Fig. 9, the performance of the electrical magnitudes can be appreciated through the power factor of the machine, which relates the active power with the total (active and reactive) power. The motor was left to rotate freely without external load torque. In steady state, the motor torque remains equal to the torque caused by the friction forces. As the power factor measures the real power consumption, it increases with the unbalance degree, because of the additional losses caused by the speed change of the shaft inertia. When running without unbalance, basically only vibrations due to friction or other mechanical causes should be appreciated. In the motor under test, a very low oscillation was detected in the balanced situation. The reason for this circumstance is the inherent unbalance present in any real machine because of the fact that the windings of the three-phase circuit are not exactly matched, but after adjusting the phase voltages, this oscillation is easily cancelled. The theoretical study indicates that the positive sequence component gives a constant torque in the forward direction, while the negative sequence component generates a second order harmonic [24]. In Fig. 10, under a 5% of voltage unbalance, we appreciate how the vibration measured has a periodic waveform (90 Hz) with a frequency that doubles the fundamental waveform, as it was expected.

Fig. 12. Speed waveforms and ILFE output amplitude for five unbalance degrees in steady state.

C. Transient and Steady-State Operation In Fig. 11, the transient response of the ILFE vibrometer is shown for a change in the supply voltages at . In this figure, only the phase is shown, because the other components amplitude remains at a constant RMS value of 200 V. The lowfrequency oscillations are caused by the sudden operation point change. In Fig. 12, the speed of the shaft was detected using a digital encoder of 4096 pulses per revolution. Using an oscilloscope with FFT module, the 90-Hz harmonic of the ILFE output was also logged for different steady-state unbalance degree. It can be seen in this figure how the speed ripple and the external vibration increase with the unbalance. In Fig. 13, this harmonic component is represented over the voltage unbalance, and as it can be seen, it exhibits a monotonic performance, increasing with the applied distortion. This means that the measurement obtained can be used in the condition monitoring of the machine to detect electrical nature faults, and in particular the ones caused by voltage unbalance. D. Harmonics Caused by Inverters In Fig. 14, the response of the ILFE sensor is represented when the power inverter injects harmonics to the motor. The


611

and the fifth (reverse rotation) and sevis 10 Hz and 50 enth (forward rotation) harmonics were created in order to emulate the performance of the proposed sensor for measuring the main components that appears in inverter-fed motors. As was shown in Section II-C, the mechanical oscillation produced can be measured at the sixth harmonic. In Fig. 15, the power inverter is overmodulated and also injects fifth and seventh harmonics. For this, the controller is demanded with a voltage higher than the maximum for the linear range (modulation index higher than the unity). The sinusoidal output of the inverter in this case tends to work as a six-step [19]. V. CONCLUSION

Fig. 13. Steady-state amplitudes of the ILFE sensor for an unbalance range of [0%–10%].

In this paper, a new comprehensive and generalized procedure is presented to predict the steady-state performance of threephase motors under unbalanced conditions. This is the first time that an ILFE fiber-optic sensor has been used as a vibrometer for unbalance detection. It is well known that voltage unbalance causes extra loads to the utilities and additional charges to consumers. With the high-sensitivity scheme proposed, specifically designed for operation in the low-frequency range, the application to voltage unbalance detection is feasible. The interferometric mechanism of the sensor allows to measure nanometric scale oscillations, with the well-known additional properties of EMI immunity and high dielectric isolation of fiber-optic sensors. These properties could make the ILFE vibrometer especially appropriate to prevent faults in high-power machines located in nuclear plants. The same sensor can also be used to detect inverter harmonics and other critical machine defects like broken bars or winding short circuits whenever they are in the sensor frequency range. ACKNOWLEDGMENT

Fig. 14. Steady-state amplitudes of the ILFE sensor measured in the sixth harmonic frequency when a fifth and seventh harmonics are injected in the voltage waveform.

The authors would like to thank the Mechanics and Physics (Electro-acoustics) Departments of the Universidad Pública de Navarra for the materials lent and G. Ghanshyam for his inestimable help. REFERENCES

Fig. 15. Steady-state amplitudes of the ILFE sensor measured in the sixth harmonic frequency when the inverter is in the overmodulation range.

digital signal processor that controls the voltages applied to the stator was reprogrammed in order to create a sinusoidal waveform superposed to the fundamental voltage. The fundamental

[1] O. Souto, J. Oliveira, and L. Neto, “Induction motors thermal behavior and life expectancy under nonideal supply conditions,” in IX Int. Conf. Harmonics and Quality of Power, Orlando, FL, 2000. [2] C. Lee, “Effects of unbalanced voltage on the operation performance of a three phase induction motor,” IEEE Trans. Energy Convers., vol. 14, no. 2, pp. 202–208, Jun. 1999. [3] I. Chindurza, D. G. Dorrell, and C. Cossar, “Vibration analysis of a switched-reluctance machine with eccentric rotor,” in Proc. 2nd Int. Conf. IEE Power Electronics, Machines and Drives Conf., Apr. 2004, vol. 2, pp. 481–486. [4] J. Undacheck and V. R. Dodd, “Progress and way out of a machinery surveillance and diagnostic system,” in ASME Conf. Pressure Vessel Piping, Mexico City, Mexico, Sep. 1976. [5] V. B. Bhavaraju and P. N. Enjeti, “An active line conditioner to balance voltages in a three-phase system,” IEEE Trans. Ind. Appl., vol. 32, no. 2, pp. 287–292, Mar. 1996. [6] T. K. Gangopadhyay, “Prospects for Fiber Bragg Gratings and fabryperot interferometers in fiber-optic vibration sensing,” Sens. Actuators A, vol. 113, pp. 20–38, 2004. [7] S. Theriault, K. O. Hill, F. Bilodeau, D. C. Johnson, and J. Albert, “High-g accelerometer based on an in-fiber Bragg Grating sensor,” in Proc. 11th Conf. Optics and Fiber Sensors, Sapporo, Japan, May 1996, pp. 196–199.

612


[8] A. D. Kersey, F. Bucholtz, and A. Dandridge, “Simple intensity-modulation based on displacement sensor with closed loop operation,” in Proc. Optical Fiber Sensors Conf., Tokyo, Japan, Oct. 1986, pp. 295–298. [9] C. Lin and F. Tseng, “A micro Fabry-Perot sensor for nano-lateral displacement sensing with enhanced sensitivity and pressure resistance,” Sens. Actuators A, vol. 114, pp. 163–170, 2004. [10] C. E. Lee et al., “Interferometric optical fiber sensors using internal mirrors,” Electron. Lett., vol. 24, pp. 193–194, 1988. [11] R. O. Claus, M. F. Gunther, A. B. Wang, K. A. Murphy, and D. Sun, “Extrinsic Fabry-Perot sensor for structural evaluations,” Applications of Fiber-Optic Sensors in Engineering Mechanics, 1993. [12] J. Sirkis, T. A. Berkoff, R. T. Jones, H. Singh, A. D. Kersey, E. J. Friebele, and M. A. Putnam, “In-line fiber etalon (ILFE) fiber-optic strain sensors,” J. Lightwave Technol., vol. 13, no. 7, pp. 1256–1263, Jul. 1995. [13] Y. Lo, S. Sirkis, and C. Chang, “Passive signal processing of an in-line fiber etalon for high strain-rate loading,” J. Lightwave Technol., vol. 15, no. 8, pp. 1578–1586, Aug. 1997. [14] J. Faiz, H. Ebrahimpour, and P. Pillay, “Influence of unbalanced voltage on the steady-state performance of a three phase squirrel-cage induction motor,” IEEE Trans. Energy Convers., vol. 19, no. 4, pp. 657–662, Dec. 2004. [15] W. H. Kersting and W. H. Phillips, “Phase frame analysis of the effects of voltage unbalance on induction machines,” IEEE Trans. Ind. Appl., vol. 33, no. 2, pp. 415–420, Mar. 1997. [16] Y. Wang, “Analysis of effects of three-phase voltage unbalance on induction motors with emphasis on the angle of the complex voltage unbalance factor,” IEEE Trans. Energy Convers., vol. 16, no. 3, pp. 270–275, Sep. 2001. [17] S. J. Chapman, Electric Machinery Fundamentals, 4 ed. New York: McGraw–Hill, 2003. [18] A. Siddique, G. S. Yadava, and B. Singh, “Effects of voltage unbalance on induction motors,” in Proc. IEEE Int. Symp. Electrical Insulation, Sep. 2004, pp. 26–29. [19] N. Mohan, T. Undeland, and W. Robbins, Power Electronics, Converters, Applications and Design. New York: Wiley, 1989, p. 328. [20] M. H. Rashid, Power Electronics, Circuits, Devices and Applications. Englewood Cliffs, NJ: Prentice-Hall, 1993. [21] S. Nandi and H. A. Toliyat, “Study of three phase induction motors with incipient rotor cage faults under different supply conditions,” in Proc. IEEE-IAS Annu. Meeting, Phoenix, AZ, Oct. 3–7, 1999, pp. 1922–1928. [22] C. M. Riley, B. K. Lin, T. G. Habetler, and R. R. Schoen, “A method for sensorless on-line vibration monitoring of induction machines,” IEEE Trans. Ind. Appl., vol. 34, no. 6, pp. 1240–1245, Nov.–Dec. 1998. [23] J. M. Corres and P. M. Gil, “Instantaneous speed and disturbance torque observer using nonlinearity cancellation of shaft encoder,” in Proc. 33rd Power Electronics Specialists Conf., Cairns, Australia, Jun. 2002. [24] C. Ong, Dynamic Simulation of Electric Machinery. Englewood Cliffs, NJ: Prentice-Hall, 1998.

Jesus M. Corres received the M.S. degree in electrical engineering in 1996 and the Ph.D. degree in 2003, both from the Public University of Navarra (UPNA), Pamplona, Spain. He has been a member of the Department of Electrical and Electronics Engineering, UPNA, for eight years, and has been involved in different projects with industry, including power systems design and motion control. His main research interests include optical fiber sensors and nanostructured materials.

Javier Bravo received the M.S. degree in electrical engineering from the Public University of Navarra (UPNA), Pamplona, Spain, in 2003. He is currently pursuing the Ph.D. degree in the Communications Doctoral program in the Electrical and Electronic Engineering Department, UPNA. His research interests are mainly in fiber-optic sensors.

Francisco J. Arregui (M’01) received the M.S. degree in electrical engineering from the Catholic University of Navarra, San Sebastian, Spain, in 1994 and the Ph.D. degree from the Public University of Navarra (UPNA), Pamplona, Spain, in 2000. He has been a member of the CEIT Research Center, San Sebastian, for two years and has been involved in different projects with industry, including medical instrumentation and monitoring of high power lines and communications hardware. Since 1995, he has been with the UPNA. During 1998 and 2000, he was a Visiting Scientist at the Fiber and Electro Optics Research Center, Virginia Polytechnic Institute and State University, Blacksburg. His main research interests include optical fiber sensors, sensor materials, and nanostructured materials. Dr. Arregui is a member of SPIE. He is an Associate Editor of the IEEE SENSORS JOURNAL.

Ignacio R. Matías (M’01-A’01-SM’03) received the M.S. degree in electrical and electronic engineering and the Ph.D. degree in optical fiber sensors from the Polytechnic University of Madrid, Madrid, Spain, in 1992 and 1996, respectively. He became a Lecturer at the Public University of Navarra, Navarra, Spain, in 1996, where hs is presently a Permanent Professor. He has coauthored more than 200 book chapters and journal and conference papers related to optical fiber sensors and passive optical devices and systems. Dr. Matías is an Associate Editor of the IEEE SENSORS JOURNAL.