AFTER an overhead distribution feeder is de-energized for

1902 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 26, NO. 3, JULY 2011 A New Technique to Detect Faults in De-Energized Distribution Feeders Part II: Symmetrical Fault Detection Xun Long, Student Member, IEEE, Yun Wei Li, Member, IEEE, Wilsun Xu, Fellow, IEEE, and Chris Lerohl, Student Member, IEEE Abstract To ensure safe re-energizing of an overhead distribution feeder after it is de-energized for an extended period, a novel fault detection technique by controlling a thyristor-based device is proposed in a companion paper. The device connected in parallel with a breaker or recloser can inject electrical pulses with adjustable strength for the downstream fault detection in a de-energized system. The proposed method can effectively detect different kinds of asymmetrical faults based on the unbalanced fault currents. However, the unbalanced current-based fault detection scheme is not effective for three-phase symmetrical faults detection. Furthermore, a stalled motor or a shunt-connected capacitor bank in the downstream may also behaves like a short-circuit. Therefore, a fault detection algorithm based on the analysis of the harmonic impedance of the de-energized system is developed in this paper. This method is very effective for the symmetrical fault detection and for distinguishing a stalled motor and capacitor bank from a fault. Extensive lab test results are provided in the paper to verify the effectiveness of the proposed method. Index Terms De-energized distribution line, fault classification, fault detection, power electronics, safe recloser. I. INTRODUCTION AFTER an overhead distribution feeder is de-energized for an extended period due to events, such as repair, maintenance, or storms, there is always the possibility that humans or animals may be in contact with feeder conductors unknowingly [1], [2]. A re-closing action in such a situation can easily lead to fatality [3], [4]. For this reason, re-energizing a de-energized distribution feeder in a safe manner is a major consideration for utility companies [5] [9]. Therefore, the development of techniques that can detect whether the downstream system is still experiencing a fault before restoring the de-energized system is important, so that operators can re-energize the feeder with confidence. The fault detection in de-energized systems is challenging since it requires the generation and injection of a voltage signal to the de-energized feeder. Available techniques for this signal generation are either self-powered which is based on battery/ca- Manuscript received October 03, 2010; revised January 07, 2011; accepted February 08, 2011. Date of publication April 05, 2011; date of current version June 24, 2011. This work was supported by the icore. Paper no. TPWRD- 00759-2010. The authors are with the Department of Electrical and Computer Engineering, University of Alberta, Edmonton, AB T6G 2V4 Canada (e-mail: xlong@ualberta.ca; yunwei.li@ece.ualberta.ca; wxu@ualberta.ca; clerohl@ualberta.ca). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TPWRD.2011.2118238 pacitor [10], [11] or by using energy from the upstream [12], [13]. A pulse-recloser technique in [14] can reduce a large inrush current; however, it is hard to distinguish a fault from a stalled motor or a shunt capacitor bank, which behaves like a short circuit. Moreover, all of the available methods cannot produce a signal with adjustable and enough strength that can be used for high-impedance fault detection. Also, these methods cannot detect all types of faults in a single device. Considering the limitations of the existing fault detection methods, a new technique based on a thyristor bridge in parallel with the breaker or recloser was proposed in a companion paper [15]. In the proposed method, a controllable signal is injected into the de-energized downstream to stimulate the electrical response. With adjustable pulse strength by changing the thyristor firing angle, a low-voltage pulse can be created to satisfy the safety requirement and a high-voltage pulse can be produced to break down an insulted gap of a high-impedance fault when necessary. The principle and key features of the proposed idea have been presented in that paper. Moreover, the method and procedure for detecting asymmetrical faults based on the unbalanced fault currents are also discussed. This paper completes [15] by proposing a symmetrical fault detection method and considering situations of downstream stalled motors or shunt capacitor banks. Since a symmetrical fault affects all three phases equally, the method based on the difference of three-phase currents for asymmetrical faults detection is not applicable here. Furthermore, when a voltage pulse is applied to the de-energized circuit, either a stalled motor or a capacitor bank can cause a large current like a short circuit. Thus, it is necessary to explore a new fault detection method that is not only based on the fault current magnitude. In this paper, a fault detection method is developed based on the harmonic impedance characteristics. Combined with the asymmetrical fault detection algorithm developed in [15], a detection scheme that can detect all kinds of faults and can distinguish a fault from a stalled motor or a shunt capacitor bank is then proposed. This paper is organized as follows. The new fault detection criterion based on harmonic impedance is introduced in Section II. In this section, the harmonic impedance characteristics under different situations are first analyzed. Based on the harmonic impedance characteristics, the proposed method for detecting a symmetrical fault and for distinguishing a fault from a stalled motor or a capacitor bank is then developed. Section III presents the lab test results obtained from a 0885-8977/$26.00 2011 IEEE

LONG et al.: NEW TECHNIQUE TO DETECT FAULTS IN DE-ENERGIZED DISTRIBUTION FEEDERS PART II 1903 TABLE I PARAMETERS OF A REPRESENTATIVE 25-kV SYSTEM Fig. 1. Proposed scheme for symmetrical faults detection. Fig. 2. Equivalent circuit with a symmetrical fault. single-phase low-voltage system, which verifies the effectiveness of the proposed method. Finally, the work is summarized in Section IV. II. FAULT DETECTION BASED ON HARMONIC IMPEDANCE As shown in Fig. 1, a zero-sequence voltage from one energized phase is fed to all three phases when the thyristors T1, T3, and T5 are fired simultaneously on a certain degree before the voltage cross zero. The corresponding current pulse in each phase depends on the line condition. In a normal condition, the injected currents are very small. However, if there is a symmetrical fault, inrush currents will show up in all phases at the same time. To illustrate this, a representative 25 kv system with a 10 km distribution line is used as shown in Fig. 2, where the parameters of the system are listed in Table I. To compare the difference of the faulted and unfaulted conditions, a symmetrical fault is created at 5 km away from the recloser. When the thyristors are fired at 150, the transient voltage and the transient current are shown in Fig. 3. Apparently, a symmetrical fault increases the current magnitude. However, this fault current magnitude highly depends on the fault resistance.it is therefore difficult to set up an appropriate current magnitude threshold to detect whether a fault exists, especially considering the possibility of a high-impedance fault. Further considering the situations of stalled motors and shunt capacitor banks connected at downstream, where a high current similar to a short circuit may be produced by the detection voltage, a fault detection method not just based on the current magnitude is required. To meet this fault detection requirement, a new method based on the harmonic impedance is introduced in this section. The harmonic impedance can represent the frequency response of power networks. With its unique characteristics, the harmonic impedance can be effectively used to detect a symmetrical fault. Fig. 3. Voltage (dv ) and current (di) in different conditions. Fig. 4. Downstream de-energized lines as a linear network. As will be shown later in this section, the harmonic impedance can also be used to distinguish a stalled motor or a capacitor bank from a fault. A. Study of Harmonic Impedance for Fault Detection Considering the downstream de-energized line as a linear network (Fig. 4), an equation can be established according to the Fourier transform of nonperiodical signals [16] where are the Fourier transforms of transient voltage and current, and is the harmonic impedance. The harmonic impedance can therefore be expressed as (1) (2)

1904 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 26, NO. 3, JULY 2011 be seen that the harmonic reactance frequency, and the ratio is is proportional to the However, if a symmetrical fault exists, the harmonic impedance will become (6) (7) where is the fault resistance. Thus, the harmonic resistance and harmonic reactance change to and Fig. 5. Harmonic resistance R(f ) in unfaulted and faulted cases. (8) (9) Fig. 6. harmonic reactance X(f ) in unfaulted and faulted cases. or where, and is the fundamental frequency (60 Hz is considered in this paper). These equations imply that the harmonic impedance can be measured by using the DFT of transient voltage and current signals. To implement this approach, the Fourier transforms of the current and voltage are created using an FFT algorithm. Separating harmonic impedance into its resistance and reactance parts gives The basic principle of the proposed idea is to utilize the change of in different conditions to detect a symmetrical fault. In normal conditions, the harmonic impedance of the de-energized downstream as shown in Fig. 2 is where of the load, (3) (4) (5) are the equivalent resistance and inductance (see Fig. 2). From (5), it can The harmonic resistance and reactance in the two conditions (faulted and unfaulted) are compared in Figs. 5 and 6, respectively. Since is much smaller than when the frequency is low, one can obtain from (8) at low frequencies (such as 0 360 Hz). On the other hand, consists of two parts: the first part is proportional to the frequency but the ratio becomes much smaller as (10) And the second part of decreases with an increase in frequency. Thus, the overall effect of the harmonic reactance is that is much smaller than and it almost has no significant change at low frequencies. Based on the above analysis, harmonic impedance can be utilized for symmetrical fault detection. The decision logic is therefore designed as follows. 1) Measure the currents and voltages of three phases after T1, T3, and T5 are fired simultaneously. 2) Are currents the same? If No, turn to asymmetrical faults analysis. If Yes, estimate the. i) If is proportional to frequency, there is no fault. ii) If has no significant change as frequency increases, there is a symmetrical fault. B. Impact of Stalled Motor A stalled motor behaves like a short circuit in the de-energized system. The equivalent circuit of a stalled motor is shown in Fig. 7 [17]. The parameters and are the motor stator resistance and leakage reactance, and the locked-rotor resistance and leakage reactance referring to the stator side are denoted as and, respectively. The reactance is the magnetizing reactance of the motor. The stator current is, and the rotor current is. The motor slip when the motor is stalled.

LONG et al.: NEW TECHNIQUE TO DETECT FAULTS IN DE-ENERGIZED DISTRIBUTION FEEDERS PART II 1905 Fig. 7. Equivalent circuit of a stalled motor. Fig. 10. Harmonic reactance X(f ) with a stalled motor. Fig. 8. stalled motor connected in the downstream system. (Note: the stalled motor: 5000 hp, starting p:f: =0:2, inrush current 700%, fed with a three-phase 5 MVA transformer, 25 kv/4.16 kv, z=0.05). Fig. 11. Shunt capacitor bank connected in the downstream. (Note: a threephase capacitor bank, 2.5 MVar, Yg connected. Fed with a 5 MVA, 25 kv/0.6 kv, Yg/Yg transformer z=0:05 pu. (12) Fig. 9. Effect of a motor on the current pulse. Considering therefore expressed as, the impedance of a stalled motor is (11) Typically, the motor starting power factor is between 0.2 0.3, and the motor inrush current is 600% 800% of the rated current. Thus, a stalled motor in a de-energized line can cause a large current pulse when a voltage is applied (Fig. 8). The currents in different conditions are illustrated in Fig. 9, in which the current pulse with a stalled motor connected is comparable to the fault current. This also indicates that the current magnitude is not a good indicator to distinguish a fault from a normal condition with a stalled motor. From the aspect of frequency domain, the impedance of a stalled motor is more inductive, and then the resistance in (11) can be ignored in a simplified model. When a stalled motor is connected parallel to a load as shown in Fig. 8, the reactance of the downstream becomes where is the reactance of downstream with a stalled motor, and is the inductance of a stalled motor. The first part of the is proportional to frequency and the ratio ; the second part of the increases with an increase in frequency. In this case, increases significantly as the frequency increases as shown in Fig. 10. This scenario is clearly different from the reactance in a faulted condition. Thus, similar to the detection of symmetrical faults, versus frequency can be used to distinguish a fault from a stalled motor. Note that using the reactance versus frequency criterion can effectively distinguish a fault from a stalled motor situation, but it cannot distinguish a stalled motor from a normal (no fault) condition as they both have a similar X/f ratio. However, detecting the existence of a stalled motor is not really the purpose of this paper. As long as a fault can be effectively detected, the proposed detection scheme can work properly. C. Impact of the Capacitor Bank As shown in Fig. 11, a shunt capacitor is usually connected in a distribution system to provide reactive power compensation and, therefore, improves the quality of the electrical supply and enhances the efficient operation of the power system. However, with its capacitive reactive power compensation, the contribution of a capacitor bank to current pulse is similar to that of a fault. When the voltage pulse is applied to a downstream with a 2.5-Mvar shunt capacitor connected, the produced

1906 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 26, NO. 3, JULY 2011 Fig. 12. Effect of a capacitor on the current pulse. Fig. 14. Harmonic impedance when a 2.5 MVA capacitor and a 5000-hp motor are connected. Fig. 13. Harmonic resistance R(f ) with a capacitor. current pulse is shown in Fig. 12. As it can be seen, the current has a comparable magnitude as the fault current. It is therefore difficult to distinguish a fault from a capacitor bank if only comparing the currents waveforms in the time domain. In the frequency domain, the capacitor almost has no contribution in the dc component (0 Hz) and its harmonic impedance decreases as the frequency increases. After installing a capacitor bank in the system, the harmonic resistance is changed (13) It is obvious that decreases rapidly with the increase of frequency. In contrast to this, the resistance in a faulted condition almost has no change in the low frequencies. When a capacitor is connected to the de-energized system, the harmonic resistance in a faulted condition and a normal condition are compared in Fig. 13. It can be seen that with the increase of frequency, the harmonic resistance decays rapidly when there is a capacitor bank. However, if a fault exists, the change of harmonic resistance in low frequencies is very small. Thus, this difference provides a criterion for distinguishing a fault from a capacitor bank. Furthermore, if a stalled motor and a capacitor bank exist in the downstream, the harmonic impedance will not follow the patterns discussed before since parallel resonance may occur with the motor and the capacitor. For example, if the motor in Fig. 8 and the capacitor in Fig. 11 are both connected to Fig. 15. Harmonic impedance when a 0.25 MVA capacitor and a 5000 hp motor are connected. the de-energized system in parallel, the harmonic impedance will become as shown in Fig. 14. Apparently, the parallel resonance occurs at the frequency 180 Hz. The magnitude of the impedance at resonance frequency is limited due to the existence of the resistance. Generally, the resonance frequency depends on the system capacitance and reactance. If the capacitance is small, the resonance will occur at a higher frequency. Fig. 15 shows the simulation result when the 2.5 MVar capacitor is replaced by a 0.25 MVar capacitor. Within the frequencies between 0 360 Hz, the scenario of parallel resonance is not observed. However, the difference between the faulted and the unfaulted cases is significant in Figs. 14 and 15. Thus, a symmetrical fault can still be detected even though there is a parallel resonance introduced by a capacitor and a stalled motor. Finally, with the aforementioned analysis for harmonics impedance characteristics under different situations, and combined with the asymmetrical fault detection method in [15], the fault detection procedure including the harmonic impedance analysis can be updated as follows. 1) Measure the currents and voltages of three phases after T1, T3, and T5 are fired simultaneously. 2) Are currents the same?

LONG et al.: NEW TECHNIQUE TO DETECT FAULTS IN DE-ENERGIZED DISTRIBUTION FEEDERS PART II 1907 Fig. 17. Diagram of the low-voltage lab test setup. Fig. 16. Overall logic for symmetrical fault detection. If No, turn to asymmetrical faults analysis (as in [15]). If Yes, calculate,, and by using the Fourier transforms of the current and voltage with an FFT algorithm. i) If harmonic resonance is observed in, it indicates that a capacitor and a motor are connected and the line is healthy. ii) Otherwise, if has a large increase with the increase of a frequency and is almost proportional to the frequency, it indicates that a normal condition or a stalled motor exists. iii) Otherwise, if decays as frequency increases, a capacitor is connected without fault. iv) If both and have no significant change as frequency increases in the observed frequencies (0 360 Hz), it indicates that a symmetrical fault exists. The overall logic for symmetrical fault detection based on harmonic impedance is summarized in Fig. 16. III. LAB EXPERIMENT A lab experiment based on a single-phase low-voltage system has been carried out to verify the proposed harmonic impedance-based fault detection method. The laboratory prototype consists of: 1) a thyristor-based signal generator; 2) a NI-DAQ based data-acquisition system; and 3) a lumped model-based equivalent circuit. Since the injected current is zero sequence in the proposed scheme, a single-phase system can effectively represent a balanced three-phase system for the research of symmetrical fault detection. In the DAQ device, six channels are utilized three of them are for voltage measurement and the other three are for current measurement. The sampling rate in the voltages and currents measurement is 1024 points every cycle, and a 1024-point FFT is performed for harmonic impedance analysis based on 60 Hz fundamental. The sampling rate is fast enough in this paper since the signal frequencies of interest are between 0 360 Hz. In some cases, such as high-impedance fault detection, higher frequencies components are also considered as will be discussed later. Due to the limitations of lab equipment, the transmission lines, transformers, and loads are replaced by equivalent R-L models. The effects of an induction motor, which is connected in parallel to the load, are also investigated. Moreover, to test the performance of the proposed idea on the high-impedance fault detection, a tree branch and a box of dirt are used to simulate high-impedance faults. The parameters of the components in this test are listed as follows, which is scaled down from the computer simulation model (with the same per-unit values preserved): power source: 120 V; transformers: 0.03 p.u., 0.76 mh; signal generator: a thyristor; firing angle is adjustable from 170 to 150 ; feeders:, 0.4 mh; load:, 10 mh; faults: resistance. The equivalent circuit of the lab test is shown in Fig. 17. The fault resistance is adjustable. Three voltage probes and three current probes measure the voltages and current at different locations. The harmonic resistance and reactance are calculated by the measured and. Figs. 18 and 19 show the and in three different conditions: 1) a bolted fault; 2) a fault with resistance ; and 3) no fault. It is seen that with a bolted fault is almost zero at all frequencies, which can be easily detected. On the other hand, increases slowly in both faulted situations but increases fast in unfaulted situations, which is consistent with the theoretical analysis. To prove that the proposed idea has the ability of distinguishing a fault from a stalled motor, a single-phase induction motor is connected to the test circuit and the parameters of the motor are listed in Table II. Since the induction motor is just like a short circuit when it is stalled, it causes a high current which is even larger than the fault current as shown in Fig. 20. To distinguish the stalled motor from a fault, in the two conditions are illustrated in Fig. 21. It is apparent that the reactance of the downstream with a motor increases obviously

1908 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 26, NO. 3, JULY 2011 Fig. 18. Comparison of harmonic resistance R(f ). Fig. 21. Harmonic reactance X(f ) with a motor. Fig. 19. Comparison of harmonic reactance X(f ). TABLE II PARAMETERS OF THE INDUCTION MOTOR Fig. 22. High-impedance fault test with a tree branch. Fig. 23. Current waveforms in a tree branch test. Fig. 20. Effect of an induction motor on current waveforms. in a normal condition, while this reactance with a fault is less affected by the increase of frequency. In reality, a fault could occur on any ground condition, like a tree branch, sand, mud, and dirt. The ground conditions affect the characteristics of faults, and changes the fault resistance. Particularly, the high fault impedance is a challenge for fault detection using the electrical signal. To investigate the performance of the proposed idea under a high impedance fault, two experiments have been carried out. One is with a tree branch (Fig. 22). The tree branch is cut off from a live tree and its length is limited to 10 cm to lower the impedance. However, its resistance still reaches 16 k, even though it is wet. Fig. 23 shows the currents obtained from the tree branch test. The resistance of the branch is too high so that the fault current is almost zero. Another high-impedance fault test is with a box of dirt (Fig. 24). The dirt comprises mud, little rocks, dead leaves, and water. The resistance is 5 k, which is lower than the tree branch but is still a high-impedance fault condition. The fault current through a box of dirt is also very low as shown

LONG et al.: NEW TECHNIQUE TO DETECT FAULTS IN DE-ENERGIZED DISTRIBUTION FEEDERS PART II 1909 Fig. 24. High-impedance fault test with a box of mud and dirt. Fig. 26. Harmonic reactance X(f ) in the high-impedance fault tests. IV. CONCLUSION Fig. 25. Current waveforms in a mud and dirt test. in Fig. 25. In this case, most of the injected current flows through the connected loads rather than the box of dirt with high impedance. The harmonic reactance in the high-impedance fault tests is shown in Fig. 26. Since there is no electrical response on the tree branch, the measured harmonic reactance in the faulted line and the unfaulted line has almost no difference. The main reason for the tree branch test not being effective is due to the low-voltage test limitation. At higher voltage levels in a typical distribution system, a stronger voltage pulse could cause actual fault arcs through a tree branch, which essentially reduces the fault impedance and, therefore, helps the high-impedance fault detection. To the contrary, the mud and dirt measurements had better results despite the fault current s low magnitude. The harmonic reactance of the faulted line follows a flatter trend than the unfaulted line within a range of frequencies, such as between 480 840 Hz. The distinction would likely improve in an actual situation, considering that the dirt had a much higher resistance than the rest of the scaled down circuit. Compared to the tree branch test, we can see that the harmonic impedance becomes more effective if there is current flowing through the fault. Once again, the situation can be expected to improve at a higher voltage level due to the arcs produced by the voltage detection signal. This paper focuses on the development of a symmetrical three-phase fault detection technique in a de-energized system, where the existence of a stalled motor or shunt capacitor bank makes the fault detection even more challenging as they behave like short circuits. Compared to the asymmetrical fault detection, the aforementioned situations require a detection technique other than simply comparing the three-phase currents. Thus, a new detection method based on the harmonic impedance characteristics under the downstream circuit is developed. The proposed method can be effectively used for symmetrical fault detection and for distinguishing a fault from a stalled motor or capacitor bank. Finally, the fault detection procedure, including symmetrical and asymmetrical faults and with consideration of possible stalled motor or capacitor banks at downstream, is presented. The proposed method has been verified in computer simulations and lab tests. This paper is supplementary to a companion paper [15], which presents the basic principle of the proposed power-electronics-based fault detection technique and discusses the procedure of asymmetrical faults detection. Combining the findings in both papers, all different types of faults in the de-energized system can be identified. REFERENCES [1] V. L. Buchholz, M. Nagpal, and J. B. Neilson, High impedance fault detection device tester, IEEE Trans. Power Del., vol. 11, no. 1, pp. 184 190, Jan. 1996. [2] S. P. Ahn, C. H. Kim, R. K. Aggarwal, and A. T. Aggarwal, An alternative approach to adaptive single pole auto-reclosing in high voltage transmission systems based on variable dead time control, IEEE Trans. Power Del., vol. 16, no. 4, pp. 676 686, Oct. 2001. [3] M. M. Eissa and O. P. Malik, A novel approach for auto-reclosing EHV/UHV transmission lines, IEEE Trans. Power Del., vol. 15, no. 3, pp. 908 912, Jul. 2000. [4] C. G. Wester, High impedance fault detection on distribution systems, in Proc. 42nd Rural Elect. Power Conf., St. Louis, MO, 1998, pp. C5 1 5. [5] IEEE Guide for Automatic Reclosing of Line Circuit Breakers for AC Distribution and Transmission Lines, IEEE Std. C37.104-2002, Apr. 2003.

1910 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 26, NO. 3, JULY 2011 [6] G. D. Rockefeller, C. L. Rockefeller, J. L. Linders, K. L. Hicks, and D. T. Rizy, Adaptive transmission relaying concepts for improved performance, IEEE Trans. Power Del., vol. 3, no. 4, pp. 1446 1458, Oct. 1988. [7] M. M. Eissa and O. P. Malik, A new digital directional transverse differential current protection technique, IEEE Trans. Power Del., vol. 11, no. 3, pp. 1285 1291, Jul. 1996. [8] J. F. Witte, S. R. Mendis, M. T. Bishop, and J. A. Kischefsky, Computer-aided recloser applications for distribution systems, IEEE Comput. Appl. Power, vol. 5, no. 3, pp. 27 32, Jul. 1992. [9] A. T. Johns, R. K. Aggarwal, and Y. H. Song, Improved techniques for modeling fault arcs on faulted EHV transmission systems, in Proc. Inst. Elect. Eng., Gen., Transm. Distrib., 1994, pp. 148 154. [10] A. T. Trihus, Cable tester, U.S. Patent 4 254 374, Mar. 3, 1981. [11] D. A. Rhein and L. R. Beard, Fault Detection System Including a Capacitor for Generating a Pulse and a Processor for Determine Admittance Versus Frequency of a Reflected Pulse, U.S. Patent 5 650 728, Jul. 22, 1997. [12] D. D. Wilson and Hedman, Fault Isolator for Electrical Utility Distribution Systems, U.S. Patent 4 370 609, Jan. 25, 1983. [13] R. Maier and W. Roethlingshoefer, Method and Arrangement for Detecting Short-Circuit in Circuit Branches of Electrical Power System Networks, U.S. Patent 5 345 180, Sep. 6, 1994. [14] S&C, S&C IntelliRupter PulseCloser: Outdoor distribution 15.5 kv and 27 kv, S&C Descriptive Bull. 766-30, Mar. 22, 2010. [15] X. Long, W. Xu, and Y. Li, A power electronics based fault detection technique in a de-energized distribution system for safe recloser operation, IEEE Trans. Power Del., submitted for publication. [16] W. Wang, E. E. Nino, and W. Xu, Harmonic impedance measurement using a thyristor-controlled short-circuit, Inst. Eng. Technol.. Gen., Transm. Distrib., vol. 1, no. 5, pp. 707 713, Sep. 2007. [17] J. C. Gomez and M. M. Morcos, A simple methodology for estimating the effect of voltage sags produced by induction motor starting cycles on sensitive equipment, in Proc. 36th Ind. Appl. Soc. Annu. Meeting, Chicago, IL, Oct. 2001, pp. 1196 1199. Xun Long (S 08) received the B.E. and M.Sc. degrees in electrical engineering from Tsinghua University, Beijing, China, in 2004 and 2007, respectively, and is currently pursuing the Ph.D. degree in electrical and computer engineering at the University of Alberta, Edmonton, AB, Canada. His main research interests include power-line signaling, distributed generation, and fault detection. Yun Wei Li (S 04 M 05) received the B.Sc. degree in engineering degree from Tianjin University, Tianjin, China, in 2002, and the Ph.D. degree from Nanyang Technological University, Singapore, in 2006. In 2005, he was a Visiting Scholar with the Institute of Energy Technology, Aalborg University, Denmark. From 2006 to 2007, he was a Postdoctoral Research Fellow in the Department of Electrical and Computer Engineering, Ryerson University, Toronto, ON, Canada. After working with Rockwell Automation Canada in 2007, he joined the Department of Electrical and Computer Engineering, University of Alberta, Edmonton, AB, Canada, as an Assistant Professor. His research interests include distributed generation, microgrid, power converters, and electric motor drives. Wilsun Xu (F 05) received the Ph.D. degree from the University of British Columbia, Vancouver, BC, Canada, in 1989. From 1989 to 1996, he was an Electrical Engineer with BC Hydro,Vancouver, and Surrey, BC, respectively. Currently, he is with the Department of Electrical and Computer Engineering, University of Alberta, where he has been since 1996. His research interests are power quality and distributed generation. Chris Lerohl (S 08) received the B.Sc. degree in electrical engineering from the University of Alberta, Edmonton, AB, Canada, in 2009, where he is currently pursuing the M.Sc. degree in power systems. His main research interests include transmission-line signaling, distributed generation, and power system harmonics.