You are on page 1of 8

IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 29, NO. 2, MARCH 2014

873

An Autoground System for Anti-Islanding Protection of Distributed Generation

Chad Abbey, Member, IEEE, Yves Brissette, Senior Member, IEEE, and Philippe Venne, Member, IEEE

Abstract—Due to the variety of distribution generation (DG) sizes and technologies connecting to distribution networks, and the concerns associated with out-of phase reclosing, anti-islanding continues to be an issue where no clear solution exists. This paper presents an autoground approach that was proposed in the context of an IEEE working group on best practices for DG protection. A prototype system was constructed using standard distribution ap- paratus and a recloser controller, and it was tested on the utility’s distribution test line. Results show that the anti-islanding detection time is approximately a cycle longer than the delay associated with application of the autoground. Once the autoground was applied, the DG was disconnected within 1 cycle on overcurrent protection. The solution is inherently scalable, applicable to all DG types, is congurable to various reclosing practices and does not require additional equipment or settings changes at the producer’s site.

Index Terms—Distributed generation, distribution relaying, re- newable energy.

I. INTRODUCTION

T HE integration of distributed generator has grown over the past decade and some utilities have reached very high

penetration levels. Despite this experience the debate between utilities and private producers still rages on a number of tech- nical issues. Possibly the most contentious is that of anti-is- landing protection [1]–[3], where a reliable high speed commu- nication based transfer-trip scheme and a local passive approach that relies only on the measurements of the voltage waveform represent, respectively, the most conservative and most liberal approaches. The anti-islanding and line protection are the two funda- mental protection requirements that need to be met by all distributed generation (DG) installations, as detailed in the DG interconnection standards [1], [2]. Line protection consists of being able to detect all faults on the distribution feeder to which the DG is connected, while not disconnecting for faults on an adjacent feeder. Generally overcurrent relaying is sufcient to meet this requirement, although in power electronic based

Manuscript received April 09, 2013; revised August 01, 2013; accepted September 16, 2013. Date of publication October 22, 2013; date of current version February 14, 2014. Paper no. TPWRS-00435-2013. C. Abbey and Y. Brissette are with the Hydro-Québec Research In- stitute, Varennes, QC J3X 1S6, Canada (e-mail: abbey.chad@ireq.ca; brissette.yves@ireq.ca). P. Venne is with Hydro-Québec Distribution, Montréal, QC H5B 1H7, Canada (e-mail: Venne.Philippe@hydro.qc.ca). Color versions of one or more of the gures in this paper are available online at http://ieeexplore.ieee.org.

Digital Object Identier 10.1109/TPWRS.2013.2284670

generators, other strategies may be necessary due to the limited contribution to short circuits by these installations. Anti-islanding has been the subject of a number of studies [3]–[9]. These approaches can be typically divided into the fol- lowing two classications: passive approaches (using the local measurements of voltage and current, and variables derived from using these quantities, to delineate between islanding and grid connected operation) and active approaches (whereby the DG perturbs either the grid voltage or frequency, an approach intended to be benign while the grid is present, and to destabi- lize the system when the substation is open. A third approach is in fact a variant on communication based approaches, whereby using thyristor valves connected to ground, a disturbance is periodically injected at the sub- station-its presence at the DG’s location indicates a normal condition, whereas its absence is indicative of an islanded grid [10], [11]. Wang et al. have also suggested these thyristor based devices for fault identication in [12]. Similar to active islanding techniques, this approach could be criticized alone on the impact on power quality. Additionally, in noisy grids or feeders that are particularly long, the issue of nuisance tripping is an issue. This paper proposes an approach to anti-islanding protection that is based on applying a three-phase short circuit to the is- landed distribution system just prior to reclosing or re-energiza- tion. Section II provides the theory and methodology for con- struction of this utility-owned equipment. Section III presents the experimental set-up and results, and we conclude with a summary of various practical considerations.

II. THEORY AND METHODOLOGY

Anti-islanding protection is required of any distributed gen- erator connecting to the distribution network, in order to protect against the case that the DG continues to energize the feeder when the utility has opened-creating an unintentional island. An unintentional island, although rather unlikely in real life, could be created by one of two scenarios. The rst case is a result of the inadvertent opening of the sub- station feeder breaker/recloser or one of the protection devices further down the feeder. This could be done in error or as a planned operation where the utility personnel do not realize that there is a DG present on the line. Eventually the line is re-en- ergized and the risk of out-of-phase reclosing exists, if the DG remains online. The second, even less likely situation would be a temporary fault that leads to operation of the utility protection device but the DG’s protection does not operate before the self-clearing fault extinguishes, creating the temporary island. For example,

0885-8950 © 2013 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

874

IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 29, NO. 2, MARCH 2014

TABLE I

COMPARISON OF ANTI-ISLANDING PROTECTION ALTERNATIVES

874 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 29, NO. 2, MARCH 2014 TABLE I OF A
874 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 29, NO. 2, MARCH 2014 TABLE I OF A

Fig. 1. Description of the autoground concept.

a tree branch might touch the line and it is cleared at the moment the vacuum bottle interrupter of a recloser operates. Although less likely, the latter of the two cases is generally used to dene the anti-islanding requirements, which links the requirement to rst period reclosing time of the local utility (2 s is given in IEEE 1547, the typical rst operation reclosing time of many North American utilities, although some are as fast as 0.5 s). This denes the speed at which the DG’s anti-islanding protection must detect the island. Table I provides a comparison of the different types of anti-is- landing approaches and the implications in terms of equipment, the installation site, relative cost, and pros and cons of each. The solution proposed in the present work, termed an autoground, is seen as a compromise in terms of cost and performance between the transfer-trip and local passive measurements. Fig. 1 presents the autoground concept, where the autoground system is installed just downline of the utility protection de- vice (substation breaker or inline recloser). In this conguration, following opening of the utility breaker, the autoground opens the substation side device, denoted sectionalizing switch (SS) and closes the autogrounding switch (AS) effectively applying a three phase to ground fault. All DGs that have not already dis-

connected based on their anti-islanding protection will be forced to disconnect based on line protection. Here it is assumed that the DG’s line protection has been properly congured in order to detect all faults, as required by [1]. In the case of inverter based DG, overcurrent protection alone may not be sufcient and more advanced functions such as overcurrent with voltage restraint (51 V) may be required. However, the focus of the present work is on the validation of the concept. Following application of the autoground for the predened time, it then ips states, opening AS following by closing of its SS. The utility breaker then re-energizes the system, without a risk of out-of-phase reclosing.

  • A. Distribution Apparatus

The autoground system consists of three main components:

the sectionalizing switch; the autoground switch, and the con- troller, which can be implemented using a variety of recloser or breaker controls. Here we describe each of these components in turn. The SS is required only to mirror the state of the substation breaker. It is preferable to have the SS as a separate device for

ABBEY et al.: AN AUTOGROUND SYSTEM FOR ANTI-ISLANDING PROTECTION OF DISTRIBUTED GENERATION

875

ABBEY et al. : AN AUTOGROUND SYSTEM FOR ANTI-ISLANDING PROTECTION OF DISTRIBUTED GENERATION 875 Fig. 2.

Fig. 2. Autoground sectionalizing switch.

ABBEY et al. : AN AUTOGROUND SYSTEM FOR ANTI-ISLANDING PROTECTION OF DISTRIBUTED GENERATION 875 Fig. 2.

Fig. 3. System autogrounding switch apparatus: assembly (left) vacuum bottle switch and ground connections (right).

the simple reason that costs escalate when work within the sub- station is required. Any savings associated with integration of the SS function into the feeder breaker would be outweighed by the cost of linking it with the AS. In this case, it is not required to interrupt fault current so any device that can provide automated sectionalizer capabilities would be sufcient. For applications where the autoground is paired with an in-line recloser, the SS is not required as its functionality can simply be integrated into the recloser itself. Fig. 2 illustrates the

ABBEY et al. : AN AUTOGROUND SYSTEM FOR ANTI-ISLANDING PROTECTION OF DISTRIBUTED GENERATION 875 Fig. 2.

Fig. 4.

Sequence of events in the operation of the autoground.

ABBEY et al. : AN AUTOGROUND SYSTEM FOR ANTI-ISLANDING PROTECTION OF DISTRIBUTED GENERATION 875 Fig. 2.

Fig. 5.

Autoground sectionalizing switch closing and opening logic.

vacuum bottle based recloser used in the experimental set-up at the base of the pole. This circuit was used to integrate outputs

as the SS. As the AS is connected in parallel with the distribution net- work only in order to apply the fault, it does not need to interrupt

from the SEL-351 relay, installed at the base of the SS pole, and used to control both the SS and the AS. The SEL-351 was chosen for the experimental set-up but any

fault current either. As a result, the apparatus is even simpler, relay used for control of distribution protection devices would

and was realized by a slight modication to an automated ca-

likely sufce. Again, in the case that the inline recloser served

pacitor bank assembly. The same vacuum bottle switches were also for the SS functionality, its controller would only need to be used but with their secondary connected to the neutral conductor reprogrammed to output signals for the AS, installed on the ad- rather than capacitor banks (Fig. 3). In this case, the switches jacent pole. The next section denes the logic required in order were actuated by relays, controlled by a simple circuit housed to realize the autoground system’s functionality.

876

IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 29, NO. 2, MARCH 2014

876 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 29, NO. 2, MARCH 2014 Fig. 6. Autoground closing

Fig. 6. Autoground closing and opening logic.

  • B. Control Logic

The sequence of events for an action of the autoground is illustrated in Fig. 4. Within this sequence, the control needs to perform three separate functions: coordinate opening of the sectionalizing switch with the upline protection device; appli- cation of autogrounding switch for a predetermined duration; and closing of the sectionalizing switch, again coordinated with the upline device. Implicit is an understanding of the reclosing strategy in place, as the design is meant to respond to the un- intentional islanding cases created during a protection device action. Fig. 5 illustrates the connection of the controller to the net- work and the logic for opening and closing of the sectionalizing switch. It is proposed to detect the opening of the upline protec- tion device using an undercurrent relay. However, detecting a sufciently small undercurrent on the three phases could not be done directly in the relay logic. There- fore, two alternatives were considered: using the complement of an overcurrent relay [NOT(5 0)] or convert the current to a voltage and use an undervoltage relay of the resulting signal. The latter was chosen as the 50 does not permit settings of less than 5 A on the medium voltage level (it is congured to mea- sure large currents). Thus, the output of the CT was connected through an appro- priate burden (to convert the current to a voltage signal) to the controller external inputs (Fig. 5). Applying these signals to an undervoltage relays (27) then enables detection of the undercur- rent condition. The sequence of events is described here. Upon opening of the upline breaker, an undercurrent is de- tected on the three phases (dened as 10% of the minimum load current on the line), signaling an opening operation of the SS. Even in the presence of distributed generation, zero current will never occur with the breaker closed, due to the inductive com- ponent of the line. Once the logic signal LT9 returns to zero (representing opening of the autoground switch, see Fig. 6) the

sectionalizing switch can be reclosed, either upon presence of voltage or by a manual close operation. In the experimental set-up, the manual close was replaced simply by a counter, cor- responding to just before the reclosing operation. Finally, the autogrounding switch is controlled by two logic circuits that then output the closing and opening pulses to two of the outputs, Fig. 6. These outputs are connected to the actuation circuit, installed at the base of the AS on the adjacent pole. As

can be noted, the trip signal from the SS causes LT9 to change state, initiating a number of delayed step functions (SV3, SV4, SV8, and SV10). The combination of these steps through two ip-ops leads to two pulses of 30 cycles, at 100 and 600 cycles (1.67 and 10 s) from the opening of the upline breaker. These two pulses are respectively used to close and open the AS. Obviously, each of these times is congurable in order to be compatible with the utility’s reclosing practices and the discon- nection time of the different DGs. The application time of the

autoground,

876 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 29, NO. 2, MARCH 2014 Fig. 6. Autoground closing

, must be compatible with the fault detection time

of all DGs for a bolted fault (symmetrical or asymmetrical) at the autoground’s location. For example, if the protection study demonstrates that the DG will trip within 6 cycles for this case,

must be longer than these 6 cycles. Otherwise, some form of voltage check must be integrated into the logic to ensure that all DGs have indeed disconnected prior to removal of the autoground.

876 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 29, NO. 2, MARCH 2014 Fig. 6. Autoground closing
  • C. Other Considerations

The previous section detailed the realization of the auto- ground system; however, there are a number of considerations prior to integration into a particular utility’s system. We discuss some of the most important ones here. The autoground system is there to serve as back-up anti-is- landing protection to the DG’s passive settings; however, rst and foremost it needs to be compatible with the distribution pro- tection practices already in place. Logically, application of the

ABBEY et al.: AN AUTOGROUND SYSTEM FOR ANTI-ISLANDING PROTECTION OF DISTRIBUTED GENERATION

877

autoground is only required for the rst reclosing period, as the DG is required to remain disconnected for 5 min after tripping. However, if different reclosing times are used and it is decided to apply the autoground during each period, the delays within the above logic need to be compatible with each of the reclosing periods. An autoground will be installed at each protection device (substation breaker, recloser) that is upline of the furthest DG along the feeder. For example, the DG might be installed down- line of two in-line reclosers, requiring three autogrounds (one for the substation breaker, and one for each recloser). In certain cases, if possible, it may be preferable to have a device moved in order to simplify coordination and limit the number of auto- ground systems that need to be installed.

With regards to the autoground itself, the vacuum bottle switches are very reliable (rated for 1200 operations) but the design still needs to handle the case that it may fail to open. This problem is taken care of by the selection of an appropri- ately rated fuse that would clear the phase that failed to operate upon reclosing of the system. Subsequent operation of the autoground would then apply an asymmetrical fault, until the issue was identied and the fuse replaced. The selection of the fuses must be done in order to coordinate with the substation protection so that the fuse would blow before the substation protection operates.

The time for which the autoground is applied,

ABBEY et al. : AN AUTOGROUND SYSTEM FOR ANTI-ISLANDING PROTECTION OF DISTRIBUTED GENERATION 877 autoground is

, needs to

take into account the line protection tripping time for all DGs. Additionally, the fuses need to be appropriate selected so that the fuses will not blow when the fault is fed from the DGs. Obviously if a DG remains online until the operation of the

autoground, it will be subjected to a fault and the resulting electromechanical stresses. This could be met with resistance from the DG proponent; however there are two arguments to address this issue. The rst is that the application of the autoground can be timed to be towards the end of the reclosing period, providing ample time for the DG to detect the islanding condition and disconnect. Secondly, the DG needs to be de- signed in order to handle these types of fault anyways so it is a bit of a non-issue, although the frequency of faults may be slightly increased.

III. EXPERIMENTAL RESULTS

  • A. Experimental Set-up

The testing was conducted on IREQ’s distribution test line, congured according to Fig. 7. The autoground system was in- stalled at the end of the rst feeder. The synchronous gener- ator was connected to the second feeder through a 600 V/25 kV transformer, as indicated. The parameters of the synchronous generator and the overcurrent protection are provided in the Ap- pendix. Switches on feeders 2 and 3 were opened in order to feed them both from feeder 1. The testing consisted of rst synchronizing the generator, bal- ancing its output power with the 150 kW load and then initiate opening of the SS for three different congurations of the auto- ground: with all three fuses in service; with one fuse removed; and with two of the three fuses removed.

ABBEY et al. : AN AUTOGROUND SYSTEM FOR ANTI-ISLANDING PROTECTION OF DISTRIBUTED GENERATION 877 autoground is

Fig. 7. IREQ distribution test line conguration for autoground testing.

TABLE II

RESULTS FOR SYMMETRICAL AND ASYMMETRICAL OPERATION OF THE AUTOGROUND

ABBEY et al. : AN AUTOGROUND SYSTEM FOR ANTI-ISLANDING PROTECTION OF DISTRIBUTED GENERATION 877 autoground is

Synchronized data was acquired using the LX-TEAC data acquisition system, monitoring the following parameters: three- phase currents and voltages at 25 kV (point 1), 600 V (point 2) and the generator torque (point 3). Additionally, measurement of the ground voltage at 1 m, 5 m and 75 m (taken to be the zero reference) was performed in order to investigate the earth voltage rise during operation of each of the cases.

B. Results

The results for the three cases are summarized in Table II. The magnitudes of the peak current (which includes the DC compo- nent) are given, along with step voltage measured at 1 meter, and the neutral current. Torque measurements were taken but they are not presented as the reading saturated in each case at 45 N m. As can be noted, the magnitude of the phase current is roughly equivalent for the three cases; however, the neutral current is much higher, as expected, for the unbalanced autoground cases. This is reected in the step voltage measurement, showing that it is around 3.5 volts for the unbalanced cases. Compared with

878

IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 29, NO. 2, MARCH 2014

878 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 29, NO. 2, MARCH 2014 Fig. 8. Substation voltages

Fig. 8. Substation voltages (top) and currents (bottom) during opening of SS and application of the AS.

878 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 29, NO. 2, MARCH 2014 Fig. 8. Substation voltages

Fig. 10. Three phase voltages (top) and currents (bottom) of synchronous gen- erator during opening of SS and application of the autoground.

878 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 29, NO. 2, MARCH 2014 Fig. 8. Substation voltages

Fig. 9. Current waveforms of the synchronous generator during application of the autoground.

878 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 29, NO. 2, MARCH 2014 Fig. 8. Substation voltages

Fig. 11. Torque measurement on the synchronous generator shaft during test (expressed on the base of the real power prior to the islanding event).

the limit of 174 volts for wet soil dened by the Electrical Code [13], the measured value represents no issue. Figs. 8–11 illustrate the case for balanced operation of the autoground, that is, with the three fuses in operation. The SS is opened at roughly 14.2 s and the AS is closed 100 cycles later (1.67 s) (Fig. 8). As can be noted, there is little change in the magnitudes of the voltages and currents of the synchronous gen- erator due to the fact that load and generation are well balanced (Fig. 10).

The application of the autoground provokes a fault current between 10 and 20 times the generator’s prefault value, leading to isolation of the generator within a cycle, Fig. 9. Given that the generator is installed right next to the autogrounding system, this represents the worst case in terms of stress on the machine. This fault current induced a small signal on the substation cur- rent measurement, which can be noted in Fig. 8. In terms of the stress on the generator, it is clear that the fault current translates to a large instantaneous torque, as can be ob- served in Fig. 11. As mentioned, the torque measurement satu-

ABBEY et al.: AN AUTOGROUND SYSTEM FOR ANTI-ISLANDING PROTECTION OF DISTRIBUTED GENERATION

879

TABLE III

SYNCHRONOUS GENERATOR PARAMETERS

ABBEY et al. : AN AUTOGROUND SYSTEM FOR ANTI-ISLANDING PROTECTION OF DISTRIBUTED GENERATION 879 TABLE III

TABLE IV

SYNCHRONOUS GENERATOR PROTECTION DEVICES AND SETTINGS

ABBEY et al. : AN AUTOGROUND SYSTEM FOR ANTI-ISLANDING PROTECTION OF DISTRIBUTED GENERATION 879 TABLE III

rated in the three cases, but the negative swing is of the order of

2.5 per unit (from 1 pu to

ABBEY et al. : AN AUTOGROUND SYSTEM FOR ANTI-ISLANDING PROTECTION OF DISTRIBUTED GENERATION 879 TABLE III

pu). This suggests that the peak

instantaneous torque would be around 3.5 pu, assuming sym- metry about the 1 pu value. Although the risk of damage to the generator may appear sig- nicant, it can be managed through various approaches. Firstly, the magnitude of the fault current-and consequently the shaft torque-can be reduced by introducing an impedance between the generator and the autoground. This could be achieved by either not installing synchronous generator based distributed generators directly at the autoground site (thereby introducing the impedance of the line), or by incorporating an impedance into the autoground system, such as connecting the output of the vacuum bottle switch to the neutral through a transformer. Secondly, the risk can be reduced statistically by lengthening the delay associated with application of the AS to just prior to reclosing. We have selected 1.67 s, in part to coincide with the IEEE 1547 requirement of 2 s for anti-islanding detection. Opening of the AS should correspond to the reclosing time (at Hydro-Québec 10 seconds is used for systems with DG). By se- lecting a delay that is just prior to reclosing would provide the DG with the maximum time for islanding detection, through its passive settings. Thus, in the Hydro-Québec system, an auto- grounding closing time at 9.5 s instead of 1.67 s, would greatly limit the cases where the DG is still online when the autoground is operated.

  • C. Recommended Practices

The above tests validated that the concept works and that the autoground system can be constructed with readily available distribution apparatus. However, many practical considerations are necessary prior to deploying the solution. First and foremost would be analysis of the impact of the au- toground technology on the local distribution protection philos- ophy. This would include ensuring that the autoground is com- patible with the number and duration of reclosing operations. This could mean adjusting the various delays mentioned above, or in certain cases, requiring different delays for different re- closing operations. That being said, it is the authors’ opinion

choices selected in the present article were made based on those

vendors Hydro-Québec has experience with but other options would also be possible. The inclusion of fuses in the design is highly recommended. Although failures of vacuum bottles are rare, they do happen and the redundancy that the fuse offers is a small additional cost.

IV. CONCLUSIONS

This paper has presented a scalable, low cost approach for anti-islanding protection of distributed generation, using a utility owned and operated system referred to as an autoground. The design of the system and its control were presented and a prototype system was constructed using standard distribution utility equipment. Testing results were presented for symmet- rical and asymmetrical operation of the system. In all cases, the DG’s overcurrent protection isolated the generator from the system in less than two cycles. The peak torque observed exceeded momentarily around 4 times the pre-fault value. The results validated the concept for a single generator based on a synchronous machine, while future work will study its applicability to systems with large numbers of DG, including those with power electronic interfaces. The solution shows promise due to its relatively low cost and inherent scalability.

APPENDIX

Tables III and IV provide information on the synchronous generator parameters and local protection devices. For the pur- poses of testing, only overcurrent protection was implemented, as the interest was in analyzing performance of the generator in the event that the island was not detected. The overcurrent pro- tection was implemented using an SEL-351 that controlled the shunt trip of an electromechanical relay. The overcurrent set- ting was set to be particularly sensitive, only slightly above the base current. In all cases, the electromechanical relay tripped the unit, evident from the shorter tripping time.

ACKNOWLEDGMENT

The authors would like to thank G. Ouellet and S. Mercier

that the autoground should only need to be applied during the for the contributions to the realization of the autoground system

rst reclosing period.

and its testing. In addition, the authors would like to thank D.

To avoid equipment issues, it would be best to consult man- Williston and other member of the IEEE 1547.8 Taskforce on

ufacturers regarding selection of the autoground apparatus. The protection for their input.

880

IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 29, NO. 2, MARCH 2014

REFERENCES

[1] Requirements for the Interconnection of Distributed Generation to the Hydro-Québec Medium-Voltage Distribution System [Online]. Avail- able: http://www.hydroquebec.com/transenergie/fr/commerce/rac- cordement_distribution.html [2] “IEEE application guide for IEEE std 1547, IEEE standard for inter- connecting distributed resources with electric power systems,” IEEE Std. 1547.2–2008, pp. 1–207, Apr. 2009. [3] J. Mulhausen, J. Schaefer, M. Mynam, A. Guzman, and M. Donolo, “Anti-islanding today, successful islanding in the future,” in Proc. Pro-

[4]

tective Relay Engineers, , Mar. 29–Apr. 1, 2010, pp. 1–8. R. J. Best, D. J. Morrow, D. M. Laverty, and P. A. Crossley, “Syn- chrophasor broadcast over internet protocol for distributed generator synchronization,” IEEE Trans. Power Del., vol. 25, no. 4, pp. 2835–2841, Oct. 2010.

[5] F. Katiraei, A. Foss, C. Abbey, and B. Strehler, “Dynamic analysis and eld verication of an innovative anti-islanding protection scheme based on directional reactive power detection,” in Proc. 2007 IEEE Elect. Power Conf., Canada, Oct. 25–26, 2007, pp. 183–188. [6] C. Abbey and P. Venne, “Field validation of a reverse reactive power relay for anti-islanding protection of a synchronous generator,” in Proc. CIGRE Canada Conf. Power Syst., Sep. 2011.

[7] Vieira, J. C. M. Vieira, W. Freitas, W. Xu, and A. Morelato, “Efcient coordination of ROCOF and frequency relays for distributed genera- tion protection by using the application region,” IEEE Trans. Power Del., vol. 21, no. 4, pp. 1878–1884, Oct. 2006.

[8]

K. El-Arroudi and G. Joos, “Data mining approach to threshold settings

of islanding relays in distributed generation,” IEEE Trans. Power Syst., vol. 22, no. 3, pp. 1112–1119, Aug. 2007. [9] S. R. Samantaray, K. El-Arroudi, G. Joós, and I. Kamwa, “A fuzzy rule-based approach for islanding detection in distributed generation,” IEEE Trans. Power Del., vol. 25, no. 3, pp. 1427–1433, Jul. 2010. [10] W. Wang, J. Kliber, G. Zhang, W. Xu, B. Howell, and T. Palladino, “A power line signaling based scheme for anti-islanding protection of distributed generators—part II: Field test results,” IEEE Trans. Power

Del., vol. 22, no. 3, pp. 1767–1772, Jul. 2007. [11] W. Wang, J. Kliber, and W. Xu, “A scalable power-line-signaling- based scheme for islanding detection of distributed generators,” IEEE Trans. Power Del., vol. 24, no. 2, pp. 903–909, Apr. 2009. [12] W. Wang, K. Zhu, P. Zhang, and W. Xu, “Identication of the faulted distribution line using thyristor-controlled grounding,” IEEE Trans. Power Del., vol. 24, no. 1, pp. 52–60, Jan. 2009. [13] C22.1–12 Canadian Electrical Code,22nd ed. 2012, Table 52.

Chad Abbey (M’01) received the B.Sc. degree in electrical engineering from the University of Alberta, Calgary, AB, Canada, in 2002. He received the M.Eng. degree and Ph.D. degrees from McGill University, Montréal, QC, Canada, in 2004 and 2009, respectively. His current research interests include wind energy, distributed generation, energy storage, and advanced distribution system automation. Since 2009, he has worked at the Hydro-Québec Research Institute (IREQ) in the area of integration of smart grid technologies and distributed generation. Among other responsibilities, he forms part of the IREQ distribution test line team and conducts research on various aspects related to the integration of renewable energy and energy storage. Prior to joining Hydro-Québec, he worked as a research engineer at CanmetENERGY, from 2004 to 2009, in a related eld. Dr. Abbey is an active member of the IEEE and CIGRE.

880 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 29, NO. 2, MARCH 2014 R EFERENCES [1] Requirements

Yves Brissette (M’81–SM’10) received the B.Tech. degree in electrical engineering from École de tech- nologie supérieure of Université du Québec, Mon- tréal, QC, Canada, in 1981. Currently, he works at the Institut de recherche d’Hydro-Québec (IREQ), Varennes, QC, Canada, on Smart grids, medium voltage breakers, switches, superconducting current limiters, and capacitive coupling systems. He also worked on a synthetic test circuit for high-voltage breakers. He is mostly working with the 25 kV medium voltage test line. Mr. Brissette is a member of the Ordre des Ingénieurs du Québec.

880 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 29, NO. 2, MARCH 2014 R EFERENCES [1] Requirements

Philippe Venne (M’00) received the B.Eng. degree in computer engineering from Université du Québec à Hull, Canada, in 2002 and the M.Sc. degree from Université du Québec à Rimouski, Canada, in 2006, where he worked on self-excited induction machine control for off-grid wind turbine applications. He is also currently pursuing the Ph.D. degree at Aalborg University in Denmark on decentralized wind farm control in a weak distribution grid. In 2009 he joined Opal-RT technologies in Mon- tréal developing models for real-time simulation of power systems. Since 2010, he has been in charge of interconnection standards for distributed generation at Hydro-Québec Distribution. Mr. Venne is a member of the Ordre des Ingénieurs du Québec.

880 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 29, NO. 2, MARCH 2014 R EFERENCES [1] Requirements