You are on page 1of 10

IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 27, NO.

2, APRIL 2012

459

Overcurrent Protection for the IEEE 34-Node Radial Test Feeder


Hamed B. Funmilayo, Student Member, IEEE, James A. Silva, Student Member, IEEE, and Karen L. Butler-Purry, Senior Member, IEEE

AbstractThe IEEE 34-node radial test feeder is a benchmark circuit for validating results from existing and novel load-ow algorithms. This paper discusses the addition of overcurrent protection (OCP) to the test feeder using off-the-shelf protective devices. The OCP scheme includes a recloser near the substation and fuses at the laterals. These OCP devices represent the conventional types present in most distribution systems. DIgSILENT Power Factory 13.1 was used to conduct the load-ow analysis, customize the OCP devices, and perform overcurrent protection studies on the test feeder. Recloser-fuse and fuse-fuse coordination tests were conducted. Results from the coordination studies showed that the assigned OCP devices provided adequate trip coordination for all fault types on the test feeder. The short-circuit results and overcurrent protective device data provided in this paper may serve as an extension to the test feeders existing load-ow data. Index TermsLoad-ow analysis, overcurrent protection, power distribution lines, power distribution protection, software tools.

I. INTRODUCTION

HE IEEE 34-NODE radial test feeder belongs to a group of existing radial test feeders for distribution system analysis [1]. Typically, the major use of these radial test feeders is to provide load-ow data which serve as a reference for validating load-ow results from existing and novel load-ow algorithms intended for radial distribution systems. Over the past few decades, the need for increased reliability and reduced cost of power delivery at the distribution level has resulted in the addition of distributed generation (DG) to radial distribution circuits [2][8]. Given the occurrence of these additions to radial distribution systems, there is a need to extend the use of the radial test feeders for various studies in order to assess the impact of DG on radial distribution systems. These studies can include overcurrent protection (OCP) studies which require nominal current from the load-ow studies and short-circuit current. However, the IEEE 34-node test feeder does not presently provide short-circuit data. For this reason, it would be necessary

to add short-circuit results to the test feeder to include overcurrent-protective (OCP) devices. The addition of these data to the test feeder will enhance the comparison of various short-circuit calculation algorithms and facilitate further studies regarding overcurrent protection for radial distribution systems. Several authors have addressed various OCP studies for radial distribution systems. For example, the authors in [2][4], [7], and [9] provided OCP schemes using example radial systems with DG while [5] and [9][16] discussed OCP for additional components found on distribution systems, such as capacitors. The IEEE 34-node radial test feeder can serve as a common circuit on which comprehensive OCP studies that are specic to radial distribution systems are assessed. In this paper, a conventional OCP scheme is proposed for the IEEE 34-node radial test feeder. OCP devices are assigned at strategic locations of the feeder. The coordination data that will allow for easy comparison and assessment of results from the overcurrent protection scheme are also provided. In Section II of this paper, the IEEE 34-node radial test feeder and the procedure for selecting the OCP devices are described. This section involves an initial comparison of the steady-state load ow and short-circuit results, from DIgSILENT PowerFactory 13.1, to the reference results of the test feeder. The load ow and short-circuit study results pertinent to the OCP study are also given. In Section III, discussions on the coordination studies (in terms of temporary and permanent faults) for various fault situations are presented. The nal list of the selected OCP devices for the test feeder is also provided in this section. Finally, the conclusions are given in Section IV. II. METHODOLOGY A. IEEE 34-Node Test Feeder System The IEEE 34-node radial test feeder system shown in Fig. 1 is a standard circuit that models the unbalanced nature observed in distribution systems (i.e., the circuit consists of one-phase and three-phase lines). The system operates on a 2.5-MVA base with a 69/24.9-kV substation transformer rating and a 24.9/4.16-kV step-down transformer rating at one of the downstream laterals. Table I shows the specications for the step-down transformer (XFM-1) at lateral 5, and the substation transformer (not illustrated in Fig. 1). The circuit consists of six spot loads and 19 distributed loads modeled as constant power, impedance, and current loads. The spot loads are all three-phase loads while the distributed loads consist of 12 one-phase, 3 two-phase, and 4 three-phase loads. Further information regarding the systems loads are provided in Tables IX and X. The systems feeders

Manuscript received April 04, 2008; accepted November 12, 2008. Date of publication March 06, 2012; date of current version March 28, 2012. This work was supported in part by the U.S. National Science Foundation under Grant ECS-02-18309. Paper no. TPWRD-00792-2007. The authors are with the Electrical and Computer Engineering Department, Texas A&M University, College Station, TX. 77843 USA (e-mail: hamed.funmilayo@centerpointenergy.com; james_silva@selinc.com; klbutler@ece.tamu.edu). 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/TPWRD.2012.2186181

0885-8977/$31.00 2012 IEEE

460

IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 27, NO. 2, APRIL 2012

Fig. 1. IEEE 34-node test feeder system (modied from [17]).

TABLE I TRANSFORMER PROFILES FOR THE IEEE 34-NODE TEST FEEDER

devices were then determined after load-ow and short-circuit studies were performed. The IEEEs short-circuit result was not directly available in [17], so it was obtained via email from Prof. W. H. Kersting. References [9], [13], [17], [19], [21], and [22] addressed fault impedance values for obtaining the minimum fault current level at the main and laterals. One reason for applying a minimum fault with some fault impedance value was to distinguish between a minimum and maximum fault on some of the single-phase laterals present in the circuit such as laterals 1, 3, and 4. The short-circuit studies were performed using a 20- fault impedance value based on reasons provided in [9] and [22]. The data utilized in providing coordination for the IEEE 34-node radial test feeder include the timecurrent curve (TCC) database of the protective devices, the maximum nominal current, as well as the minimum and maximum steady-state fault currents (shunt and branch) from the laterals and the main feeder. Only one node was faulted at any point in time for each fault scenario observed. B. Steady-State Analysis In analyzing the system protection at the distribution level, the load-ow and short-circuit results are essential for determining the ratings for the protective devices to be employed in the system. The IEEEs model circuit in [17] uses an ideal balanced voltage source which is directly connected to node 800 at 1.05 p.u. and 0/-120.0/120 , for the load-ow analysis. For the short-circuit studies, the IEEEs model circuit assumes a voltage source at 1.0 p.u. with the substation transformer applied between the voltage source and node 800. Therefore, the load-ow and short-circuit studies results were determined using two circuits. In this paper, the load-ow and short-circuit studies were implemented on a single model circuit. The reason is because the short-circuit studies method present in DIgSILENT initially calculates the prefault load ow. This condition implies that the prefault load-ow results must match the load-ow results. The following subsections present a detail of the load-ow and short-circuit analysis as implemented in this paper. 1) Load-Flow Data: The load-ow study of the IEEE 34-node radial test feeder was implemented in DIgSILENT to match the recommended prole at node 800 (1.05 p.u. per phase). In order to match this prole, the substation transformer taps in DIgSILENT were adjusted and xed at 12, 10, and 9. The voltage prole obtained on using the xed substation transformer taps were 1.05/1.046/1.047 p.u. with angles 0.002/-119.463/120.808 for phases a, b, and c, respectively. The taps could be adjusted in 32 steps from 16 to 16. The negative orientation of the tap value indicates an increase in the voltage prole at the controlled node (node 800). Table II presents the load-ow results comparing the IEEEs maximum nominal branch current through the OCP devices to those of DIgSILENT. Two terms From Node and To Node are used in expressing the branch currents orientation. The From Node is the node that the nominal branch current ows out from and the To Node is the node that the nominal branch current ows into. The percentage error difference between the branch current results from the IEEE and those from DIgSILENT are located in the last column of the table. The error values were calculated using the IEEE results as the base. The

were modeled in DIgSILENT using the conductors 4 4 primitive impedance matrix, which were derived from the 3 3 unbalanced phase impedance matrix provided in [17]. As observed in Fig. 1, the test feeder includes components, such as the capacitor banks, a step-down transformer, as well as the dened main feeder and laterals. These components all require some form of overcurrent protection. There are a total of 6 one-phase laterals labeled 1, 2, 3, 4, 6, and 8; and 4 three-phase laterals labeled 5, 7, 8, and 11. Laterals 5 and 7 are completely three-phase lines while lateral 8 combines a three-phase line with a one-phase line. Also, laterals 5 and 7 are of special interest due to their unique characteristics with respect to the main feeder. Lateral 5, for example, operates on a different voltage level, 4.16 kV, and produces an undervoltage at its farthest node of 890. The reason for the undervoltage is mostly due to the high magnitude of the constant-current spot load and the line losses along the length of the lateral. Lateral 7, on the other hand, is the only location with reactive compensation as given in [17]. The two capacitor banks identied in [17] are dened as Cap-844 and Cap-848, and they are the extensions of nodes 844 and 848, respectively. Finally, Lateral 11, dened as the line section from node 836840, shares similar characteristics to the main feeder. However, using some engineering judgment mentioned in [18][20], the line section is considered a lateral because of its location and proximity to the other three-phase laterals with components. This situation implies that in the event of a fault at that location (836840), only that section will be isolated and the systems reliability is enhanced. 1) Overview of Procedure for Including Overcurrent Protective Devices in the Radial Feeder: The primary goal of this paper is to provide overcurrent protection for the IEEE 34-node test feeder. This study involves determining the satisfactory ratings and settings for the OCP devices in the system to accommodate different fault levels. Customized versions of the OCP device models available in PSS/ADEPT 5.3.2s protection database were modeled in DIgSILENT to conduct the protection coordination study. To begin with, the recloser was placed near the substation while a fuse was placed on each lateral. The settings of the OCP

FUNMILAYO et al.: OVERCURRENT PROTECTION FOR THE IEEE 34-NODE RADIAL TEST FEEDER

461

TABLE II MAXIMUM UPSTREAM BRANCH CURRENT FROM THE LOAD-FLOW RESULT

TABLE III COMPARISON OF MAXIMUM FAULT CURRENT TO IEEE RESULTS

TABLE IV COMPARISON OF MINIMUM FAULT CURRENT TO IEEE RESULTS

percentage error values for the branch current results obtained in DIgSILENT are within a 2% margin from the IEEE results. 2) Short-Circuit Data: As discussed earlier, the circuit model used for conducting the short-circuit studies was the same as that used for the load ow. DIgSILENTs complete method was used for the short-circuit study. This method initially conducts the prefault load ow before the short-circuit values are calculated. The transformers taps for each phase remained xed at the values discussed in Section II-B1. The voltage regulators RG10 and RG11, which are located within the circuit, were modeled without the line drop compensation parameters provided in [17]. Both regulators maintained their tap values during the short-circuit analysis and their operation is in agreement with [12]. To begin with the short-circuit analysis, the branch fault current owing through each OCP device was used for determining the operating times of the device during maximum and minimum faults. The maximum fault was applied at the nearest node downstream of the OCP device while the minimum fault was applied to the farthest node downstream of the OCP device. A three-phase-to-ground fault was used as the maximum fault for the three-phase lines while a single-line-to-ground fault without impedance was applied at the one-phase laterals. For the minimum fault, a single-line-to-ground fault with a 20- fault impedance was used at all laterals. Tables III and IV provide the comparison of results from the IEEEs maximum and minimum shunt fault current, within the vicinity of the OCP devices, to the results generated from DIgSILENT. Results from Tables III and IV indicate that the percentage errors observed between the fault-current values are within 10% for the faulted locations. The IEEE* label shown in the tables implies that the reference short-circuit fault current values were obtained via correspondence with the author of [17] but have not been made publicly available in [17]. The percentage errors observed during the fault studies are larger when compared to the nominal branch current values. One reason for this difference is because the same circuit was used to conduct load-ow and short-circuit studies as discussed in Section II-B. Hence, the short-circuit calculation accounted for the impedances which were introduced to model substation transformer tap positions to achieve the 1.05-p.u. prole

at node 800. Another noticeable difference in Table IV is that the percentage errors were greater near the transformer locationsnodes 800 and 890. The reason is because DIgSILENT requires an extra parameter called the short-circuit voltage (%) in addition to the X/R ratio value provided. The typical short-circuit voltage value which corresponded with the respective kilovolt-amperes of the transformer was assumed from [23]. The shunt fault current magnitude near the transformers decreased as the assumed short-circuit voltage value was increased. For example, when the assumed short-circuit voltage value for XFM-1 (value of 4.5%) was approximated to 5% (0.5% increase), the shunt fault current magnitude at node 890 decreased by 14 A (3% decrease). These reasons account for the higher percentage errors at nodes 800 and 890. C. Selection of Overcurrent Protective Devices The selection of OCP devices was determined according to recommendations from [9], [10], [19], and [24]. It was also stated in [9] and [19] that momentary or temporary faults occur more often than permanent faults at the distribution system level. With this in mind, the selected recloser was set to operate prior to the fuses for faults downstream of the fuses. This operation allows the faults to clear since they temporarily and consequently save the fuses from operating unnecessarily. The previously described process called fuse-saving was implemented at most of the fuse locations selected for the test

462

IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 27, NO. 2, APRIL 2012

Fig. 2. IEEE-34-node test feeder system with protective devices.

feeder. Fig. 2 illustrates the IEEE 34-node radial test feeder with the OCP devices assigned at strategic locations. As observed in Fig. 2, the boxes labeled F signify the fuses and the box labeled R signies the recloser. The recloser was placed on the main feeder while the fuses were placed on the laterals. In this OCP scheme, the recloser also acts as the nal backup protection for the system during a permanent fault on the laterals. The following subsections describe the requirements for the recloser and fuse types. 1) Recloser and Load-Side Fuses: The reclosers coordination range must provide adequate time to sense all faults that may occur at any node downstream of the feeder. A triple singlephase electronic recloser is preferable as opposed to a threephase recloser since most of the loads were one phase. Although the three-phase recloser has a longer reach, its operation during a fault affects all three phases of the feeder simultaneously and, therefore, provides less exibility for the loads to operate independently. The triple single-phase recloser as dened in [9] is an electronically controlled unit that can operate in three different modes, namely: 1) single phase; 2) single-phase trip with three-phase lockout; and 3) three-phase mode. These modes provide some additional exibility to the recloser operation. Based on recommendations from [9], similar types of fuse links were selected as OCP devices for all branches within the same nominal current range. The load-side fuses (fuse cutouts with fuse links) had a voltage rating equal to or higher than that of the maximum bus voltage at the fuse location. Also, the interrupting current rating was larger than the maximum symmetrical fault current at the location of the fuse. In this paper, the selected load-side fuses were of types K, T, and X expulsion fuse links with external cut outs, and they presented the most effective coordination range with the reclosers instantaneous settings. 2) Transformer Fusing: Due to the presence of a step-down transformer (XFM-1) on lateral 5, several protection options described in [9] and [15] are available in order to assign the proper protection for the transformer and the load on that lateral. The rst choice is an internal fuse link, which can be embedded within the transformer and coordinated with a secondary breaker. This option is economical but it does not allow the fuse to be eld replaceable after a fault event in the vicinity of the transformer. A second and moderately economical option is to provide an external expulsion cutout in place of the internal fuse. In this case, the fuse can be replaced after all faults, and this option is well suited for low-current faults. The last option described involves a combination of a current-limiting fuse in se-

ries with an expulsion cutout. This method would account for low and high current faults but it is rather expensive to install. Given the list of available options described in the last paragraph, the choice of protection provided for XFM-1 consisted of a type T external expulsion cutout on the transformers primary side. Some transformer fusing application principles, such as voltage rating, short-circuit interrupting rating, ampere rating, and the speed characteristic of the fuse were considered. The voltage rating of the fuse must be equal to or greater than the voltage at the transformer location. As for the ampere rating, the value should be the same or exceed the anticipated normal loading level for the transformer. Similarly, the symmetrical short-circuit interrupting rating of the transformers primary fuse should match or exceed the maximum possible fault current at the transformer location. Second, the transformer primary fuse must be capable of withstanding the inrush currents generated whenever the transformer is energized. The fuse must also protect against transformer faults and secondary-side faults (through faults). Proper assessment of transformer faults requires information about the transformer impedance, kilovolt-ampere rating, and the nominal voltage. Similarly for assessing the secondary-side fault currents, knowledge of the transformer winding conguration is needed. For any secondary fault condition, the value of the fault current observed at the transformer secondary winding differs from the fault current observed at the transformers primary side. This difference is highly dependent on the transformer connection type. The symmetrical (wye-wye) connection across the transformers winding lends itself to a 11 p.u. ratio for evaluating the fault current at the transformer primary [13][15]. The selected fuse must then coordinate with the upstream recloser to maintain fuse savings. Moreover, the fuse must coordinate with the OCP device downstream of the lateral to serve as its backup device. This protection option should provide adequate sensitivity for the faults on lateral 5. 3) Capacitor Fusing: Capacitors are sources of transient fault current owing to their mode of operation (charging and discharging) at very high frequencies [11], [14]. However, in the event of a failed capacitor unit (at steady-state frequency), the type of protection fuse required must promptly isolate the failed capacitor unit on the line prior to the operation of any other protective device on the system. The rating for such a fuse is dependent on the connection type of the capacitor unit, the operating voltage, and the maximum line-to-ground fault current at the capacitor location [9], [11], [14]. In the IEEE 34-node test feeder, the units of the capacitor bank are arranged in a grounded wye-wye conguration. According to[9] and [11], the current through each capacitor unit on each phase is rated at 135% of the nominal phase current. This overrating is necessary in order to prevent the capacitor fuses from being damaged by inrush currents that may occur during capacitor bank switching and lightning surges. Also, for the worst case scenario of a capacitor unit failure, a one-phase grounded fault current without fault impedance is assumed as the capacitor fault value [11]. Two fusing methods are available for protecting the capacitor banks as discussed in [9]. These methods are group fusing and individual fusing. Individual fusing involves one fuse protecting

FUNMILAYO et al.: OVERCURRENT PROTECTION FOR THE IEEE 34-NODE RADIAL TEST FEEDER

463

TABLE V MINIMUM FAULT CURRENT OBSERVED AT THE RECLOSER FOR THE MINIMUM FAULT AT EACH LATERAL

each capacitor unit, and this method is mostly implemented for substation capacitor banks. In group fusing, one fuse protects the capacitor bank. This method is regarded as more economical and common for capacitor banks located downstream of the feeder, and was therefore implemented in this paper. An additional step that was not applied in this study was to coordinate the selected fuse link with the capacitors tank rupture curve. The reason for not implementing this additional step was because the tank rupture curve specications for the capacitors were not provided in [17]. D. Determination of Settings The settings for the protective devices are determined after the appropriate types of protective devices have been selected. Two objectives are required regarding the choice of protective device settings. Foremost, the device must not operate unnecessarily and second, it must provide adequate sensitivity to all fault types by interrupting the fault in a timely manner. The basic rule of thumb in ensuring maximum coordination between a recloser and lateral fuses is to set the recloser curves to operate without interrupting the operation of the fuse. The sequence for the recloser was set for one instantaneous and two delayed trip operations (1-fast, 2-delayed) before locking out. Although the 2-fast, 2-delayed trip is widely used, the selected sequence for the recloser helps reduce fault duty on the transformers in the system [9]. Also, since the recloser type is electronically controlled, an additional time delay can easily be allocated following its instantaneous trip, thereby eliminating the need for a second instantaneous trip operation [9]. The following paragraphs state each protective devices rating while keeping the coordination objectives in mind. 1) Recloser: The reclosers minimum trip rating was chosen to exceed the maximum nominal branch current through the recloser, as shown in Table II. Hence, in order for the recloser to operate for temporary faults and provide fuse savings, the fault current sensed by the recloser during the minimum fault at any lateral must exceed the reclosers minimum trip rating. Table V provides values of the branch fault current ( recloser) that will

be the sensed by the recloser when a minimum fault (1-ph fault with impedance) is applied at each lateral (faulted node). The nominal voltage for the selected electronic recloser type was 14.4 kV L-N while the minimum trip rating was 100 A. Table V shows that the recloser coordinates with the fuses at most locations even at the farthest end of the main feeder since ( recloser) exceeds the reclosers minimum trip rating at these locations. Based on these specications, the reclosers operating range is from 100 to 12 000 A. The selected reclosers curve types for the reclosers instantaneous and delay trip sequences were 103 and 134, respectively. In this paper, the letters A and BB have been used to identify the electronic reclosers instantaneous and delay trip curve types, as a means of simplicity. Fuses on certain locations of the circuit will not be coordinated with the recloser to achieve the fuse-saving concept discussed in Section II-C. For example, the fuse between nodes 888890 at lateral 5 will not be coordinated with the recloser during a minimum fault at node 890. One reason is because the recloser and the fuse are not operating at the same voltage level. Other locations which recloser-fuse coordination will not be applicable are Cap-844 and Cap-848 on lateral 7. When a fault occurs at the capacitors locations, it is preferable that the recloser not operate on its instantaneous trip curve (A) prior to the complete operation of the fuses protecting the capacitors [9], [10]. This condition will prevent possible overvoltages generated by the capacitor bank in the event of a capacitor fault. Further discussions on these unique cases are presented in Section II-E.3. 2) Load-Side Fuse Sizes: The load-side fuses voltage ratings as well as the nominal and symmetrical current ratings were selected from PSS/ADEPTs protection database and customized in DIgSILENT. The voltage ratings were matched to the 24.9 kV L-L and 4.16 kV L-L system rating, while the nominal current rating of each fuse was based on the branch current value at each fuses location. The laterals, which had similar nominal current ranges, were grouped, and the fuse size that matched the current range of each lateral group was selected. This approach is intended to simplify the fuse replacement needs required during future scheduled maintenances [9]. The processes of determining the ratings for the transformers fuse and the capacitors fuses were not as straightforward and required further calculations. 3) Transformer Fuse Sizes: The transformers (XFM-1) primary fuse specications were determined based on discussions in subsection C2. According to [15], since the rated voltage of the transformer exceeds 600 V and the impedance is less than 7%, then the maximum-allowable fuse rating for the XFM-1s fuse is 300% of the full-load current. XFM-1s voltage and impedance ratings have been presented in Table I of this paper. According to [15], secondary-side (through) faults are considered to be the most difcult to interrupt by the transformer-primary protection devices. The method for calculating the through faults are provided in [13]. These fault currents must be interrupted by the selected primary fuse. Using the rated voltage, the calculations for the transformer faults are given in (1). The should be sensed by the calculated primary fault current

464

IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 27, NO. 2, APRIL 2012

selected transformer-primary fuse and interrupted in a timely fashion (1) where % 4) Capacitor Fuse Sizes: The calculations involved in determining the shunt capacitor fuse ratings for nodes 848 and 844 are provided in (2) and (3). The grounded shunt capacitor at of 24.9 kV with 450-kVAR node 848 has a nominal voltage reactive power (Q). Since the capacitor bank is wye connected, is calculated as its rated current (2) According to [9], the fuse link (F9) should have a minimum of 1.35 times the rated current as this factor accounts rating for currents that could be generated by harmonics. Therefore, applying the factor to the rated current generated in (2) provides a new , which is calculated in (3) as (3) Similarly, the grounded shunt capacitor at the node 844 has a of 24.9 kV with a total of 300-kVAR reacnominal voltage tive power (Q). Therefore, the fuse link should have minimum current rating calculated as 9.45 A. Finally, the ratings selected for fuses F9 and F10 should satisfy coordination requirements with Fuse F7 at the lateral end for fuse-fuse coordination. E. Coordination Studies Now that the OCP devices have been selected for the system, the coordination studies follow. This study is implemented as a screening process for selecting the nal list of OCP devices for the test feeder. The study primarily ensures that the optimal protective devices selected and applied to strategic locations on the feeder do operate correctly during a fault. In this paper, two terms are used in describing the OCP devices to be coordinated. These terms are the primary and secondary (backup) device. The primary device is closer to the fault and is the rst to clear the fault if it is permanent. The secondary device is a backup of the primary device in case the primary device fails to clear the fault. The coordination scheme employed TCCs of the recloser and fuses to address the impact of temporary and permanent faults on device coordination. Ideally, for temporary faults, the reclosers instantaneous trip should operate completely before the fuses minimum melt begins to operate in order to enhance proper fuse-saving operation [9]. Recloser-fuse coordination was implemented for temporary fault conditions. Based on [9], factor to the reclosers instantaneous the application of a trip is required in order to assess the recloser-fuse coordinafactor is a multiplier whose tion for temporary faults. The value is dependent on the reclosers sequence of operation. The factor was applied to the recloser time values in order to provide a safety margin between the reclosers operation and

the minimum melting time of the downstream fuses that are factor value coordinated with the recloser. In this paper, a of 1.25 was selected. This value corresponded to the reclosers 1-fast, 2-delayed operation selected in Section II-D. On the other hand, for a permanent fault situation at the lateral, the fuse must completely operate prior to the reclosers delayed trip operation. In other words, the maximum clearing time of the fuse must be below the reclosers delayed trip time. This condition will ensure that the recloser functions properly as a backup protection in case the fuse fails to interrupt the fault. As for the locations where no fuse saving was required, a derating factor concept was introduced to implement a fuse-fuse coordination between the primary fuse and the fuses backup. According to [9] and [19], the derating factor of 75% sufces for effective fuse-fuse coordination. Applying the factor implies that the maximum clearing time of the primary fuse will not exceed 0.75 times the minimum melting time of the secondary fuse. This measure ensures that the primary fuse operates completely before the backup fuse. Failure to implement these coordination basics may result in fuse fatigue and minimize the fuses dependability. With respect to the overcurrent protection scheme designed for the IEEE 34-node radial test feeder, ve possible fault cases will be described with respect to temporary and permanent fault conditions. Highlights of the OCP devices operation and coordination (recloser-fuse and fuse-fuse) during each case will be discussed. 1) Fault on the Main Feeder: The simplest case requiring protection is that of a fault on the main feeder at the 24.9-kV level. For this scenario, the recloser is the only OCP device required to operate on its instantaneous trip within two cycles to clear the fault given that the fault is transient or temporary. If the fault is not cleared, it becomes a permanent fault. In this case, the reclosers delayed trip will operate twice then lockout since no fuse is located on the main feeder. 2) Fault on Ordinary Laterals: In this case, for any fault on laterals 1, 2, 3, 4, 6, 8, and 11, the recloser operates on its instantaneous trip to clear the fault if the fault is temporary, thereby enhancing fuse savings. If the fault is not interrupted and remains on the lateral, then the respective fuse operates to clear the fault and isolates the lateral from the rest of the system. An example is presented in Fig. 3 to show the protective devices operations during a minimum fault occurrence at node 810 on lateral 1. In the plot, the recloser interrupts the fault and trips on its instantaneous curve A at 0.056 s. If the fault is not cleared, the fuse (F1) interrupts the fault and begins to melt at 0.091 s. After 0.901 s, the recloser will trip on its delayed curve BB, to serve as a backup in case the fault was not cleared by the fuse. 3) Fault on Laterals With Reactive Compensation: In this case, two scenarios will be described. The rst involves a fault on any of the nodes present on Lateral 7 excluding the capacitor nodes. In this event, the recloser trips on its fast curve for a fusesaving operation. If the fault is not cleared, fuse F7 interrupts and clears the fault, after which the lateral is isolated. The two capacitor fuses are not involved in this sequence of events. The second scenario describes a capacitor fault. This situation is represented by a bolted single line-to-ground fault at either of

FUNMILAYO et al.: OVERCURRENT PROTECTION FOR THE IEEE 34-NODE RADIAL TEST FEEDER

465

Fig. 3. Recloser-fuse coordination for the minimum fault at 810.

Fig. 5. Fuses F5 and F12 operation for a maximum fault on node 890.

Fig. 4. Operation of fuse F10 during a capacitor fault at node Cap-848.

the capacitor nodes 844 or 848 (Cap-844 and Cap-848). The capacitor fuse at the faulted zone will be the rst protective device to interrupt the fault. Fuse F7 on the lateral serves as the backup device to isolate the entire lateral before the reclosers delayed trip operation begins. One important distinction of the second scenario is that the recloser still instantaneously trips prior to the operation of the backup device (F7). This fuse-saving operation is not desired if the fault is to be promptly cleared by the backup device. This situation provides more justication for choosing the (1-fast, 2-delayed) recloser operation instead of the (2-fast, 2-delayed) option. Fig. 4 illustrates the capacitor fuse (F10) operation at Cap-848 during a capacitor fault. As observed from the gure, the capacitor fuse F10 clears the fault in approximately three cycles (0.044 s). The recloser will operate on its A curve after four more cycles have elapsed in case fuse F10 does not clear the fault. Fuse F7, the backup device for F10, will then begin to melt at 0.380 s in order to isolate the lateral. 4) Fault on Laterals With Step-Down Transformer: This case describes protective device coordination for faults downstream of XFM-1. Recloser-fuse coordination was applied to the recloser and the XFM-1s primary fuse (fuse F5) while fuse-fuse

coordination was applied to fuse F5 and load-side fuse, F12. A typical example describing these coordination types occurs during a fault on the farthest node on the lateral (node 890). In this situation, fuse F12 operates to clear the fault and fuse F5 serves as its backup. For a fault upstream of fuse F12, for example, on node 888, the recloser operates instantaneously for fuse saving and serves as the backup for fuse F5 in case the fault becomes permanent. Since fuse F5 and the recloser operate at the same voltage level, fuse saving is possible during the minimum and maximum faults. Fig. 5 illustrates the fuse-fuse coordination plot for F5 and F12 during a maximum fault on node 890. This protection scheme would protect the load downstream on node 890 and the transformer during temporary or permanent faults. As observed from Fig. 5, fuse F5 operates at 24.9 kV while fuse F12 operates at 4.16 kV. During a maximum fault near the load (node 890), fuse F12 senses a fault current of 440.122 A and operates at 0.110 s on its minimum melt and 0.260 s on its maximum clear. If the fault is not cleared, fuse F5 (backup device) operates at 0.913 s on its minimum melt and 1.353 on its maximum clear while sensing a fault current of 73.533 A. III. DISCUSSION OF RESULTS A. Coordination Studies From the description of the four cases in the last section, the rst two of the four cases of the coordination studies, that is, Section II (E1 and E2), would fully benet from the fuse-saving operation. Table VI highlights the recloser-fuse coordination time intervals of the protective devices during the minimum and maximum fault, as observed from DIgSILENT. The results were computed by comparing the reclosers instantaneous time to the fuses minimum melting time. The coordination time interval was calculated based on the difference in the operating times of both devices. As observed from the time interval results in the table, the coordination time interval between the recloser and fuses exceeds 1 cycle or more during the maximum faults (worst case) with a few exceptions. These exceptions are fuses 9 and 10.

466

IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 27, NO. 2, APRIL 2012

TABLE VI RECLOSER-FUSE COORDINATION TIME INTERVALS FROM DIGSILENT

TABLE VIII OVERCURRENT PROTECTIVE DEVICES LIST FOR THE TEST FEEDER

B. Selected OCP Devices Studies All 12 fuses dened in Table VIII were standard fuse types. These selected fuses and the recloser were available in the PSS/ ADEPT protection database. As observed from Table VIII, six of the seven KEARNEY manufactured fuses (rated at 4 A) were selected to protect the lightly loaded laterals. The six fuses include fuses 1, 3,4,6,8, and 11; these fuses provided the appropriate level of sensitivity for the protected laterals. Furthermore, given that the loads in the system were assumed as static loads, the effect of future load changes on the ratings of the selected OCP devices were assessed by considering long-term load growth. The typical longterm (110 years) load growth rate assumed for OCP design ranges between (13)% [19], [25]. References [4] and [19] discussed the calculation of load growth using parameters, such as the load growth rate (g%), the initial and nal value of the and ), and the number of years considered ( th load ( year). This calculation is given in (4) and the growth rate and the number of years were varied in order to address the impact of increasing load changes on the selected OCP devices ratings provided in Table VIII (4) From (4), the selected OCP devices used in the work and reported in this paper were capable of accommodating an annual load growth rate of 1.25% over a ve-year period. This statement implies that the ultimate load currents may be increased by a factor of 1.06 without causing any violations to the settings of the selected OCP devices provided for the system. The selected fuses presented in the table provided the maximum coordination range with the recloser when the system was faulted. Moreover, the maximum symmetrical current rating of shown in the table exceeded the all OCP devices maximum fault present on the system. For this reason, the OCP devices can be said to accommodate higher fault current levels if they occur within the system. IV. CONCLUSION A conventional OCP and coordination scheme has been successfully implemented on the IEEE 34-node radial test feeder using DIgSILENT PowerFactory 13.1. DIgSILENT has the functionality of modeling the unbalanced characteristic

TABLE VII FUSE-FUSE COORDINATION TIME INTERVALS FROM DIGSILENT

The negative time interval values observed for fuses 9 and 10 indicate that each fuse operated on its maximum clear curve prior to the recloser instantaneous trip operation. In these excepfactor to the recloser may seem redundant tions, applying a and no recloser-fuse coordination for fuse saving is required. Fuse F5 coordinates with the recloser during a transformer fault described in Section II (C2 and D3). Results from interpolating the transformer fault current show that the recloser will operate at 0.161 s on its instantaneous curve to coordinate with fuse F5. Table VII presents the result for the fuse-fuse coordination. Observations from Table VII show that the ratio of the primary fuses maximum clearing time to the secondary fuses minimum melt time is well below the derating factor limit (0.75) required for proper fuse-fuse coordination during a minimum fault. Table VIII illustrates the nal list of selected overcurrent protective device types for the test feeder. The table provides the devices ratings and the manufacturer names.

FUNMILAYO et al.: OVERCURRENT PROTECTION FOR THE IEEE 34-NODE RADIAL TEST FEEDER

467

TABLE IX SPOT LOADS

TABLE X DISTRIBUTED LOADS

on lateral 5 during temporary or permanent (maximum and minimum) fault events on the lateral. For lateral 7, the capacitor fuses were selected to operate completely prior to the reclosers instantaneous trip operation in the event of a capacitor fault. Thus, fuse-fuse coordination was implemented between the capacitor fuses (F9 and F10) and F7 while recloser-fuse coordination was implemented between F7 and the reclosers instantaneous trip curve. Finally, the overcurrent protective devices list presented in this paper, along with the coordination studies results, may serve as reference data to extend the use of the IEEE 34-node radial test feeder as a benchmark circuit. The results will enable easy comparison and assessment of future overcurrent protection studies regarding radial distribution systems with or without additions such as DG. ACKNOWLEDGMENT The authors would like to thank F. J. Verdeja Perez, J. Mendoza, S. Duttagupta, M. Marotti, K. Manseld, T. Djokic, and H. E. Leon for their contributions, along with the assistance of Prof. W. H. Kersting. REFERENCES
[1] W. H. Kersting, Radial distribution test feeders, in Proc. IEEE Power Eng. Soc. Winter Meeting, Columbus, OH, 2001, vol. 2, pp. 908912. [2] J. C. Gomez and M. M. Morcos, Coordination of voltage sag and overcurrent protection in DG systems, IEEE Trans. Power Del., vol. 20, no. 1, pp. 214218, Jan. 2005. [3] R. C. Dugan and D. T. Rizy, Electric distribution protection problems associated with the interconnection of small, dispersed generation devices, IEEE Trans. Power App. Syst., vol. PAS-103, no. 6, pp. 11211125, Jun. 1984. [4] B. Fardanesh and E. Richards, Distribution system protection with decentralized generation introduced into the system, IEEE Trans. Ind. Appl., vol. IA-20, no. 1, pp. 122130, Jan. 1984. [5] Power System Relaying Committee Working Group, Intertie protection of consumer-owned sources of generation3 MVA or less. 1990. [Online]. Available: www.osti.gov/energycitations/product.biblio.jsp?osti_id=6719106-13k [6] P. P. Barker and R. W. D. Mello, Determining the impact of distributed generation on power systems: Part 1Radial distribution systems, in Proc. IEEE Power Eng. Soc. Summer Meeting, 2000, pp. 16451656. [7] A. Girgis and S. Brahma, Effect of distributed generation on protective device coordination in distribution system, in Power Eng. Large Eng. Syst., Halifax, NS, Canada, 2001, pp. 115119. [8] V. V. Thong, E. Vandenbrande, J. Soens, D. V. Dommelen, J. Driesen, and R. Belmans, Inuences of large penetration of distributed generation on N-1 safety operation, Proc. IEEE Power Eng. Soc., p. 5, 2004. [9] Cooper Power Systems, Electrical distribution-system protection, 2005. [10] S. R. Mendis, M. T. Bishop, J. C. McCall, and W. M. Hurst, Capacitor overcurrent protection for industrial distribution systems, IEEE Ind. Appl. Mag., vol. 2, no. 3, pp. 2027, May/Jun. 1996. [11] S&C Electric Company, Data Bull. 350-130, Selection Guide for the Protection of Overhead Distribution Capacitor Banks. Aug. 1992. [Online]. Available: http://www.sandc.com/edocs_pdfs/ EDOC_037204.pdf [12] IEEE Recommended Practice for Protection and Coordination of Industrial and Commercial Power Systems, IEEE Standard 399-1997, Aug. 1998. [13] J. Burke, Hard to nd information about distribution systems, ABB Inc. Consulting, Raleigh, NC, 2002. [Online]. Available: library.abb. com/.../VerityDisplay/4EDF68B14B79751F85256C550053D6B6/ $File/Hard-to%20Find%2019c.pdf [14] IEEE Recommended Practice for Electric Power Distribution for Industrial Plants, IEEE Standard 141-1993, Dec. 1993. [15] IEEE Guide for Protective Relay Applications to Power Transformers, IEEE Standard C37.91-2000, Mar. 2000.

of distribution systems. Load-ow and short-circuit studies were conducted on the same circuit by modeling the substation using an ideal voltage source and substation transformer with xed taps to produce a 1.05-p.u. prole at the transformers secondary. This model is more versatile for OCP studies which require load-ow and short-circuit results. The load-ow results have been validated using the reference from the IEEE Distribution System Analysis Subcommittees publication in [17]. The short-circuit and coordination studies were performed by choosing a 20- fault impedance value for the minimum faults. This choice does not imply the minimum fault current possible, it only indicates the minimum fault that the OCP devices are set to trip based on the protection philosophy. The nal list of selected OCP devices, following the short circuit and coordination studies, provided the most suitable OCP device types at their respective locations on the test feeder. Results from the coordination studies show adequate recloser-fuse coordination at all locations operating at the same voltage level on the feeder. Two solutions were proposed to address the 2 three-phase laterals with unique characteristics. An expulsion fuse F5 selected for lateral 5 was chosen to provide recloser-fuse coordination with the reclosers A curve and fuse-fuse coordination with fuse F12 , which is located downstream of the lateral. Therefore, fuse F5 provided adequate overcurrent protection for the step-down transformer and load

468

IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 27, NO. 2, APRIL 2012

[16] P. J. Meyer, Primary-side transformer protection. S&C Electric Company, Feb. 2005. [Online]. Available: www.sand.com/edocs_pdfs/ EDOC_27579.pdf [17] W. H. Kersting, R. C. Dugan, and S. Carneiro, Jr., Jan. 2001, Radial Test Feeders-IEEE Distribution System Analysis Subcommittee. [Online]. Available: http://www.ewh.ieee.org/soc/pes/dsacom/testfeeders. html [18] F. Soudi and K. Tomsovic, Optimal trade-offs in distribution protection design, IEEE Trans. Power Del., vol. 16, no. 2, pp. 292296, Apr. 2001. [19] T. Gonen, Electric Power Distribution System Engineering. New York: McGraw-Hill, 1986. [20] K. Tomsovic and F. Soudi, Optimized distribution protection using binary programming, IEEE Trans. Power Del., vol. 13, no. 1, pp. 218222, Jan. 1998. [21] J. W. Goodfellow, Understanding how trees cause interruptions, an update of ten years of high voltage research, 2006. [Online]. Available: http://www.utilityarborist.org/images/Presentations/ISA%202006%20Goodfellow.pdf [22] J. Dagenhart, The 40-ohm ground-fault phenomenon, IEEE Trans. Ind. Appl., vol. 36, no. 1, pp. 3032, Jan/Feb. 2000. [23] B. D. Metz-Noblat, F. Dumas, and C. Poulain, Calculation of shortcircuit current, 2005. [Online]. Available: http://www.schneider-electric.com/documents/technical-publications/en/shared/electrical-engineering/electrical-know-how/low-voltage-minus-1kv/ect158.pdf [24] 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. 2732, Jul. 1992. [25] H. M. Al-Hamadi and S. A. Soliman, Long-term/mid-term electric load forecasting based on short-term correlation and annual growth. 2005. [Online]. Available: http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6V30-4FWK6DN2&_user=952835&_rdoc=1&_fmt=&_orig=search&_sort=d& view=c&_acct=C000049198&_version=1&_urlVersion=0&_userid=952835&md5=44169e23ee88400a06d6ffbdfc414fb0

Hamed B. Funmilayo (S04) received the B.Sc. degree in electrical engineering from Kansas State University, Manhattan, in 2005. Currently, he is a Graduate Research Assistant at Texas A&M Universitys Power System Automation Lab. His research includes distributed generation with an emphasis on overcurrent protection. Mr. Funmilayo is a member of Eta Kappa Nu and the Power Engineering Society.

James A. Silva (S06) received the B.S.E.E. and M.S.E.E. degrees in electrical engineering from California Polytechnic State University, San Luis Obispo, in 2005, and is currently pursuing the Ph.D. degree in electrical and computer engineering at Texas A&M University, College Station. For the 20052006 academic year, he was a Teaching Assistant with the School of Electrical and Computer Engineering, Purdue University, West Lafayette, IN. His research topics have included power system protection, distributed generation (with an emphasis on wind power), and microgrids.

Karen L. Butler-Purry (SM01) received the B.S. degree (Hons.) in electrical engineering from Southern University, Baton Rouge, LA, in 1985, the M.S. degree from the University of Texas, Austin, in 1987, and the Ph.D. degree in electrical engineering from Howard University, Washington, DC, in 1994. She joined Texas A&M University, College Station, in 1994, where she is currently Professor in the Department of Electrical and Computer Engineering. Her research interests are in the areas of distribution automation and intelligent systems for power quality, equipment deterioration, and fault diagnosis. Dr. Butler-Purry is a Registered Professional Engineer in the States of Louisiana, Texas, and Mississippi. She is a member of the Power Engineering Society, the American Society for Engineering Education, and the Louisiana Engineering Society. She received the National Science Foundation Faculty Career Award in 1995 and the Ofce of Naval Research Young Investigator Award in 1999.

You might also like