406 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 22, NO. 1, JANUARY 2007 Overcurrent Protection on Voltage-Source-ConverterBased Multiterminal DC Distribution Systems Mesut E. Baran, Senior Member, IEEE, and Nikhil R. Mahajan, Student Member, IEEE Abstract—This paper proposes a protection scheme which utilizes modern voltage-source converters as fast-acting current-limiting circuit breakers. This paper investigates the main challenges of detecting and localizing a fault, and interrupting it as quickly as possible in a multiterminal dc system. A system protection scheme consisting of smart relays associated with converters has been developed. The protection relays monitor local quantities to detect and isolate disturbances/faults. It is shown that overcurrent-based schemes can be adopted for these relays to meet the fast response requirements. The effectiveness of the proposed protection scheme is illustrated through simulations. Index Terms—DC distribution, power-electronic converters, protection. NOMENCLATURE CDCCB ETO PEBB SES VSC Capacitor dc circuit breaker. Emitter turnoff device. Power-electronic building block. Shipboard electrical system. Voltage-source converter. (CBs) and converter action [6]. Protection becomes more challenging for multiterminal dc lines, as dc CBs are needed to isolate the faulted section of the system. Recently, there have been advances toward the development of dc CBs [7]–[10]. In a recent HVdc application [1], an insulated-gate bipolar transistor (IGBT)-based dc CB was utilized. Furthermore, the new semiconductor devices used in the new converters have the capability to limit and interrupt fault currents [8]. Therefore, it is possible to integrate fault-current handling capability within the new converters and have them behave similar to fast-acting current limiting CBs (i.e., limit and interrupt the fault current [11]). With this new functionality for converters, a new challenge emerges, that the protection scheme should also be able to detect and locate faults more quickly. To achieve this goal, this paper proposes a new protection system using relays embedded into the converters. The application considered is power distribution on a ship. It is shown that by adopting overcurrent-based schemes for these relays, this type of system can be protected against faults on the dc lines. It is also shown that these converter-based relays can detect and isolate faults quite fast, on the order of a few milliseconds. II. DC SYSTEM PROTECTION I. INTRODUCTION T HE emergence of voltage-source converters (VSCs) that use self turnoff power-electronic devices makes dc distribution an attractive alternative for medium- and low-voltage distribution in special applications. Present day applications include dc ties between two systems at medium-voltage levels [1], [2], and dc distribution for industrial parks, space, and shipboard electrical systems (SESs) [3], [4]. One of the main limitations of present day VSCs is that their fault current withstand is much lower than that of thyristorbased converters, typically twice the nominal current rating of the converter [5]. Hence, faults on a dc line fed by the VSCs must be limited and interrupted much faster than those on a conventional HVdc system. Protection of the dc lines requires special consideration. For a simple two-terminal dc line with two converters, protection is usually achieved by a combination of ac-side circuit breakers Manuscript received May 25, 2005; revised January 24, 2006. This work was supported by the Office of Naval Research (ONR) under Award N000014-00-10475. Paper no. TPWRD-00315-2005. The authors are with the Department of Electrical and Computer Engineering, North Carolina State University, Raleigh, NC 27695 USA (e-mail: baran@ncsu. edu; nikhilmahajan@ieee.org). Digital Object Identifier 10.1109/TPWRD.2006.877086 The basic task of the protection scheme proposed here is the same as any other protection scheme—detect and locate any disturbance that can occur on the system and isolate the affected area quickly. The prototype system considered for this study is shown in Fig. 1. It is a revised version of the system considered by the U.S. Navy for SES [12], [13], and it represents a generic multiterminal dc distribution system. The system has two ac sources (generators) feeding the dc line through rectifier converters. The boost-type voltage-controlled bridge rectifiers are rated at 4 MW each, and they convert the 4160-V, three-phase ac voltage to 7 kV dc. In this system, the dc line is a short cable (100 m). Sectionalizers have been placed on the middle of the line so that only one side of the line would be out of service during a fault on the dc line. The loads are supplied through the dc buck converters or inverters. For protection of such a system, relays and protection devices should be placed on the system to detect faults, interrupt fault current, and isolate the faulted part of the system. These design issues are addressed in the following subsections. A. Protection Devices To provide protection for system components (generators, converters, and loads), the conventional approach would be to put CBs at the input and output of the converters in order to facilitate unit protection for each device. 0885-8977/$20.00 © 2006 IEEE Authorized licensed use limited to: IEEE Xplore. Downloaded on April 8, 2009 at 14:03 from IEEE Xplore. Restrictions apply. BARAN AND MAHAJAN: VSC-BASED MULTITERMINAL DC DISTRIBUTION SYSTEMS Fig. 1. Prototype dc distribution system with CBs. The alternative approach, which has been considered here, is to use the converters themselves for fault current interruption. In [11], for example, it has been shown that VSCs used for dc distribution can limit fault current and interrupt it and, hence, they can be used as current-limiting CBs. Furthermore, converters have overcurrent protection on the power semiconductor switches to protect the switches when the current through a switch gets close to its limit. Therefore, this switch level protection employed on the converter can serve as a backup overcurrent protection relay. Hence, in this study, the CBs are assumed to be embedded within the converters. As a result, the protection zones defined by them will not be device based, and Fig. 2 illustrates the main zones defined by embedded CBs on the prototype dc system. Note that there will be four main zones. 1) DC zone: The dc line supplies power to all of the loads and, therefore, it is the most critical component for protection. It is also the one that is exposed the most to the faults/damage. Note that since the switches in the converters will be conducting the fault interruption, the protection zone is defined by the switches of the converters that are connected to the line side of the rectifiers, inverters, and the dc converters. The zone, therefore, includes not only the dc line but also the dc rails of the converters connected to the dc line, as the figure illustrates. Note that to protect the dc line itself by using conventional schemes; we would need a CB at every connection point, as illustrated in Fig. 1. The proposed scheme eliminates these extra CBs. 2) Rectifier ac side zone: A protection device is needed at the source side terminals of the rectifier to provide protection against faults on the input filter elements of the rectifier, as these elements are upstream of rectifier switches. Fig. 2 illustrates this small zone. In terrestrial applications, the protection device at the rectifier input would usually be an ac CB. However, for the prototype system considered, fuses have been chosen instead of CBs, as Fig. 2 illustrates, mainly since the fault in this protection zone will be permanent rather than temporary. Thus, there is no need for a fast reclosing capability that the CB can provide. Also, the compactness of fuses is an important advantage for the shipboard systems. 3) Inverter load-side zone: The inverters are tapped off the dc line to supply power to the ac loads. Since the inverter can provide overcurrent protection for the loads 407 it serves [14]–[17], no protection device is needed at the load side. There is no need for a CB on the source side either, since the inverter does not feed a fault on the source side assuming a nonregenerative load (except the input filter capacitor). 4) DC converter load side zone: Similar to the inverter, the dc–dc buck converter, when properly designed, can provide overcurrent protection on the load side. Furthermore, if the dc–dc converter is of the two stage type, as shown in Fig. 10, the input side switches can also provide fault isolation on the source side. Hence, with this functionality, there is no need for a protection device at either terminal of the dc-dc converter. Capacitor Protection: When a fault occurs on the dc line, the capacitors connected to the dc line (such as the filter capacitors at rectifier output, and inverter and the dc–dc converter inputs) start to discharge with a very short time constant and contribute to high fault currents. Therefore, the capacitors connected to the dc buses of converters demand special attention in the form of their own protection. Protection of the capacitors is typically done at a hardware level by way of snubbers [17] which limit the rate of discharge current. The snubbers, however, do not interrupt fault current. To limit and interrupt the discharge current during a fault, a protection device is needed. For the prototype application, a CDCCB has been employed [8] as shown in Fig. 3. Without the snubber, the time constant of the capacitor discharge current will be around 10 s. Hence, the CDCCB should also be very fast in order to effectively protect the capacitor from extreme stresses and destruction [16], [17]. In [5] and [8], it is shown that indeed an ETO-based CDCCB can be used to turn off and interrupt fault current in less than 10 s, thereby meeting the requirement set forth. The operating principle of the CDCCB is based on the inherent current sensing of the ETO [5]. The measured current is compared to a 2.1-kA threshold (maximum limit of the DCCB kA). When the capacitor current crosses this threshold, a hard turnoff is initiated which limits the current from further increase and interrupts the current in 3–7 s, thus protecting the capacitor satisfactorily. Fig. 4 illustrates the capacitor discharge current limiting using the CDCCB. This case corresponds to the discharge of the rectifier output capacitor on the prototype system in Fig. 2, and the simulation is done using EMTDC softs ware [19], [20]. As the figure illustrates, the fault at causes the capacitor discharge current to rise very rapidly, and it crosses the threshold level of 2 kA in about 4 s, at which time, the CDCCB interrupts the fault current. As the figure illustrates, this is a very effective method of discharge current limiting. Its use is warranted here for two reasons: 1) to limit the amount of energy to be dumped by the capacitor to the dc cable and 2) to limit the discharge currents that will flow through the converter switches following a fault since if they are not limited, the high discharge currents can damage the switches. B. Fault Monitoring and Protection For detecting faults and operating appropriate protective devices, a protection system consisting of relays integrated with Authorized licensed use limited to: IEEE Xplore. Downloaded on April 8, 2009 at 14:03 from IEEE Xplore. Restrictions apply. 408 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 22, NO. 1, JANUARY 2007 Fig. 2. Zones for the prototype dc system without separate CBs. Fig. 3. CDCCB protection. each VSC has been considered. The goal here is to provide autonomy to the relays so that they can take fast action. The challenge to be addressed here is to provide enough intelligence for the relays to ensure that they will make correct decisions. The details about these relays embedded in the rectifier, buck converter, and the inverter are given below. 1) Rectifier Relay: The faults on the dc line are the most severe faults on the system. In the prototype application, the primary concern is the bus fault (short-circuiting of the two terminals of the dc line) since it will cause high fault currents from the rectifiers. The smoothing capacitors at the output terminals of the rectifier will react to a dc bus fault first and start discharging very fast, as the bus fault will cause the dc bus voltage to collapse very fast. As pointed out in the previous section, that is why the capacitors need a very fast acting device, such as CDCCB, to limit and interrupt this high discharge current. A typical bridge-type boost rectifier (with antiparallel diodes) will also feed high current to the fault, as the rectifier will act like a diode bridge once the dc terminal voltage collapses. Hence, the rectifier diodes need to be replaced with turnoff devices in order to be able to interrupt the current before it gets too high [11]. To detect the faults on the dc side of the rectifier, the relay monitors the dc output current I_R shown in Fig. 5 and uses an Fig. 4. Capacitor discharge current during a fault and its interruption by CDCCB. (a) Capacitor discharge. (b) Zoom-in of (a). overcurrent protection scheme: if I_R passes a threshold value and stays above it for a certain amount of time, it assumes that there is a fault on the dc side of the rectifier. To improve the security of this scheme, the relay also monitors the dc bus voltage, Authorized licensed use limited to: IEEE Xplore. Downloaded on April 8, 2009 at 14:03 from IEEE Xplore. Restrictions apply. BARAN AND MAHAJAN: VSC-BASED MULTITERMINAL DC DISTRIBUTION SYSTEMS 409 Fig. 7. Fuse operation. Fig. 5. Zones around the rectifier. Fig. 6. Fault current limiting by the rectifier. (a) Generator currents. (b) Rectifier current. and checks whether the voltage drops below 80% of its rated value. When both the overcurrent and low-voltage conditions are satisfied, the relay sends a trip signal to the rectifier to interrupt the current and isolate the rectifier from the dc line. Simulation results for a dc bus fault are shown in Fig. 6. Fol, the bus capacitor discharges into lowing the fault at the fault with a very short time constant as seen in Fig. 4, and the CDCCB limits and interrupts the discharge current in less than 10 s. Meanwhile, the generator also starts contributing to the fault current as seen by the I_A, I_B, I_C, and I_R in Fig. 6. s, When I_R exceeds the threshold of 1.75 kA at the relay detects and identifies a primary dc-bus fault as seen in Fig. 6(b), as by this time, the bus voltage has already collapsed. The relay then sends a trip signal to turn off the rectifier switches in order to interrupt the fault current. Note that the fault current here is limited and brought down to zero by the rectifier switches in a controlled manner (within 20 s) in order to limit the transient recovery voltages across the switches to an acceptable level (see [11] for details). The relay also monitors the input ac line currents I_A, I_B, or I_C, to decide whether the fault is on the dc or ac side. If the fault is on the ac side, the output current I_R will not rise. Hence, by monitoring ac-side currents, the relay detects ac-side faults and turns off the rectifier switches in order to isolate the fault. Note that this action does not interrupt the ac-side fault current. The ac-side faults are interrupted by the fuse, as illustrated below. Rectifier Fuse: As pointed out above, fuses at the source side of rectifiers are needed to protect the rectifier ac zone shown in Fig. 5. The fuse provides the primary protection for the L–L faults and the three-phase faults on the ac source side of the rectifier. The fuse will not operate for ground faults, as the system is high resistance grounded. The fuse also provides backup protection for the rectifier relay. The fuse should be fast enough so that the upstream protection does not trip before the fuse blows, and it should be slow enough so that the downstream protection has enough time to operate. Since the downstream protection in this case is much faster than the operation times of a fuse, coordination of the fuse with the downstream relay of the rectifier PEBB is not a problem. For the prototype system, the current profiles are: 1) the nominal line current is about 450 A (rms), 2) the startup current is about 1.5 kA (rms) for about 5–10 ms (due to initial capacitor charging via rectifier PEBB), and 3) the fault current level is about 10–30 kA (depending on the fault being after/before the source inductor). Hence, a 500E-rated fuse is appropriate for this application, and the EJO-1 type 9F62 fuse from GE has been considered for the prototype system. To demonstrate the fuse operation, a fuse model was developed for EMTDC/PSCAD to represent the fuse melting and clearing. A phase A–B fault at the rectifier input terminals has been simulated, and the results are shown in Fig. 7. As the figure s, the line currents shows, following the fault at I_A and I_B start increasing and they are limited only by the , the fuse clears as it has source impedance. At dissipated enough energy for clearing, and at s, at the first zero crossing of I_A and I_B, the fault currents have been interrupted. Hence, the fuse interrupts the current in about half a cycle (about 10.4 ms), as desired. 2) Inverter Relay: Commonly used three-phase bridge-type inverters have been considered here for the prototype system. These inverters can be used to provide overcurrent protection for the load-side faults [11], [14]–[17] with minimal revisions. This is because, unlike the rectifier, the antiparallel diodes of the inverter do not freewheel the current during a load-side fault. Fig. 8 shows the zones of protection associated with the inverter. The inverter relay monitors the input and output currents Authorized licensed use limited to: IEEE Xplore. Downloaded on April 8, 2009 at 14:03 from IEEE Xplore. Restrictions apply. 410 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 22, NO. 1, JANUARY 2007 Fig. 8. Protection zones around the inverter. , and , , , to detect and localize the faults. Note that the relay is responsible mainly for the faults in the load zone, as the upstream faults in the source zone are interrupted by the rectifiers. To illustrate the fault current interruption by the inverter, an overcurrent scheme is considered (although other more secure schemes are possible [14]–[17]). The inverter relay monitors the input current , and when the current passes the threshold limit, the relay assumes the fault. At the same time, if any of the output currents pass the corresponding threshold value, the fault is localized as the load-side fault. Note that since the system is high resistance grounded, the first ground faults are not disruptive and, hence, only the line and three-phase faults are considered for the prototype system. Simulation results for a three-phase fault at the load terminals of the inverter are given in Fig. 9. The three-phase fault causes the output voltage to collapse and, correspondingly, the current starts to increase. When the input current crosses the threshold of 2.5 kA and stays above the threshold, the relay detects the fault and turns off the switches of the inverter to interrupt and isolate the fault. Note that the current decreases slowly to zero, rather than being chopped when the switches are turned off. This is due to the fact that the antiparallel diodes provide a freewheeling path for the load current after the turnoff. As a result, there is very little transient recovery voltage on the switches. Note also that monitoring the output currents helps the relay to localize the fault (i.e., identify whether the fault is downstream or internal to the inverter). Note that the load considered here is passive. If there is a motor load that can regenerate, a diode on the input dc bus is needed to prevent reverse current feed to a source-side fault. 3) Buck Converter Relay: dc–dc buck converters can also perform fault-current interruption when properly designed [11]. A buck converter with an isolation transformer provides enough flexibility for this purpose, and it is considered for the prototype application. The converter in Fig. 10 uses an isolation transformer with a 5:1 turns ratio to help buck the primary 7-kV dc voltage to 800 V dc. Note that the switches do not use antiparallel diodes, and this facilitates fault-current interruption and isolation by turning off all of the switches as will be illustrated here. The protection relay for the buck converter is mainly responsible for detecting and interrupting the faults in the load-side Fig. 9. Fault current limiting by the inverter. (a) Inverter current profile during a fault. (b) Output voltages. Fig. 10. Zones around buck converter. zone as Fig. 10 illustrates. The protection scheme considered for this relay is similar to that of the inverter. The relay monitors the input current , output current , as well as transformer currents and . The relay detects the existence of a crosses a threshold (0.5 kA). The relay localizes fault when the fault to the load side when the output current crosses the . threshold (2.5 kA) along with a simultaneous increase in The simulation results are shown in Fig. 11 for a bus fault at . The fault causes the curthe load-side terminal at to increase as shown in Fig. 11(a). At , rent crosses the threshold of 0.5 the relay detects the fault as kA. Also, at , the currents and cross the threshold of 2.5 kA as shown in Fig. 11(b). The relay thus identifies the fault as the load-side bus fault. The protective action taken by the relay is to turn off the converter switches. Authorized licensed use limited to: IEEE Xplore. Downloaded on April 8, 2009 at 14:03 from IEEE Xplore. Restrictions apply. BARAN AND MAHAJAN: VSC-BASED MULTITERMINAL DC DISTRIBUTION SYSTEMS Fig. 11. Fault current limiting by a dc–dc buck converter. (a) Secondary dc bus fault detection. (b) Secondary dc bus fault locating. (c) Diode freewheeling. The action of hard turnoff of the switches interrupts the current from the primary side immediately, but the current on the load side does not cease immediately. The energy stored in the output inductor is freewheeled through the diodes of the rectifier stage. When this energy is completely dissipated, the fault is completely interrupted. The freewheeling of the diodes is seen in Fig. 11(c) from onwards. This freewheeling of the diodes is important in preventing any excessive transient recovery voltages on the downstream devices [11]. C. Protection Coordination and Backup For the proposed protection scheme to provide complete system protection, the relays should provide: 1) proper coordination and 2) backup when a protection relay fails. 411 Coordination: Coordination between relays is needed to provide selectivity (i.e., only the protective device closest to a fault should operate to isolate the faulted part of the system [6]). On the prototype system, for example, coordination is needed between the rectifier relay and fuses, between the two rectifier relays, and the rectifier relays and the load converter (inverter and buck converter) relays. This coordination can be achieved by adopting the same techniques used for coordinating overcurrent protection devices as described in the following. • Coordination Between Rectifier Relay and Fuse As shown earlier, coordination between the rectifier relay and the fuse is achieved by the proper selection of the fuse. The fast action fuse selected for the rectifier takes 10–100 ms (depending on fault current) to interrupt and isolate the ac zone faults, whereas the relay detects and interrupts the dc zone faults in about 0.5 ms. Therefore, the relay coordinates with the fuse by detecting and interrupting faults in its protection zone much faster than the source-side fuse. • Coordination Between Rectifier Relays Since multiple rectifiers feed the dc line, when there is a fault on the ac side of a rectifier, the other rectifiers will feed the fault also, as the rectifier switches can carry the bidirectional currents. Hence, coordination between rectifier relays is needed so that when a fault is in the ac zone of one of the rectifiers, only the relay of the corresponding rectifier interrupts the fault. For this coordination, we propose to employ an interface diode on the dc side of the rectifier shown in Fig. 5. This diode will prevent the reverse dc current flow through the rectifier and, thus, it will eliminate the need for coordination among the rectifier relays. • Coordination Between Rectifier and Buck-Converter Relays The rectifier relay needs to coordinate with the relays of the load converters (buck converters and inverters). The coordination is facilitated by the fact that the rated current of load converters is much lower than that of the rectifiers. For the prototype SES, for example, the normal current of the buck converter is 200 A and, therefore, the threshold for its relay is set to 500 A (2.5x). Since the threshold of the rectifier relay is 1.75 kA, the big difference in thresholds provides sufficient margin for coordination between these two relays. Backup: On the prototype system, local backup protection and the remote backup protection provide protection when the primary protection fails. Remote backup is the main backup scheme, and it is based on the fact that power flows from the source (generators) toward the loads. Hence, the buck converter relay provides backup for the inverter relay, and the rectifier relay provides backup for the buck converter relay. Finally, the fuse provides backup for the rectifier, as previously pointed out. The overcurrent protection used for the switches themselves on the converters provides also a “local backup” feature which is unique to the dc system considered. This local protection can be considered as a backup relay when the primary relay associated with the converter fails. For example, if a fault occurs on the dc line of Fig. 2, the rectifier relay provides primary protection. But if the relay fails, the switch level protection on the rectifier then Authorized licensed use limited to: IEEE Xplore. Downloaded on April 8, 2009 at 14:03 from IEEE Xplore. Restrictions apply. 412 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 22, NO. 1, JANUARY 2007 provides local backup for the relay by turning off the rectifier switches when the current through the switches approaches their limits. [9] J. Zyborski, T. Lipski, J. Czucha, and S. Hasan, “Hybrid arcless lowvoltage ac/dc current limiting interrupting device,” IEEE Trans. Power Del., vol. 15, no. 4, pp. 1182–1187, Oct. 2000. [10] P. M. McEwan and S. B. Tennakoon, “A two-stage dc thyristor circuit breaker,” IEEE Trans. Power Electron., vol. 12, no. 4, pp. 597–607, Jul. 1997. [11] N. R. Mahajan, “System protection for PEBB based dc distribution systems,” Ph.D. dissertation, Dept. Elect. Comput. Eng., North Carolina State Univ., Raleigh, NC, Nov. 2004. [12] K. L. Butler, N. D. R. Sarma, C. Whitcomb, H. Do Carmo, and H. Zhang, “Shipboard systems deploy automated protection,” IEEE Comput. Appl. Power, vol. 11, no. 2, pp. 31–36, Apr, 1998. [13] J. G. Ciezki and R. W. Ashton, “Selection and stability issues associated with a navy shipboard dc zonal electric distribution system,” IEEE Trans. Power Del., vol. 15, no. 2, pp. 665–669, Apr. 2000. [14] D. Kastha and B. K. Bose, “Investigation of fault modes of voltage-fed inverter system for induction motor drive,” IEEE Trans. Ind. Appl., vol. 30, no. 4, pp. 1028–1038, Jul./Aug. 1994. [15] A. Ghosh and G. Ledwich, Power Quality Enhancement Using Custom Power Devices. New York: Springer, 2002. [16] R. Peuget, S. Courtine, and J.-P. Rognon, “Fault detection and isolation on a PWM inverter by knowledge-based model,” IEEE Trans. Ind. Appl., vol. 34, no. 6, pp. 1318–1326, Nov./Dec. 1998. [17] F. Blaabjerg and J. K. Pedersen, “A new low-cost, fully fault-protected PWM-VSI inverter with true phase-current information,” IEEE Trans. Power Electron., vol. 12, no. 1, pp. 187–197, Jan. 1997. [18] E.-C. Nho, I.-D. Kim, and T. A. Lipo, “A new boost type rectifier for a DC power supply with frequent output short circuit,” presented at the Industry Applications Conf., Phoenix, AZ, 1999. [19] Manitoba HVDC Research Centre, “EMTDC, The electromagnetic transients & controls simulation engine”. Winnipeg, MB, Canada, 2002. [20] Manitoba HVDC Research Centre, “PSCAD, power systems computer aided design” . Winnipeg, MB, Canada, 2003. [21] M. Baran and N. Mahajan, “System reconfiguration on shipboard DC zonal electrical system,” in Proc. IEEE Electric Ship Technologies Symp., Philadelphia, PA, Jul. 2004. III. CONCLUSION This paper illustrates that protection of a multiterminal dc distribution system can be simplified by utilizing the power-electronic converters’ ability to act as fault current-limiting CBs. It is shown that by associating relays with the converters—rectifier, inverter, and dc–dc buck converter—and by adopting overcurrent-based protection schemes for these relays, the faults on the system can be detected and localized very fast. The simulation results on the prototype system illustrate that these relays can quickly detect and localize the faults within a few milliseconds. The relays in the proposed scheme are versatile in that they can perform both unit protection and overcurrent protection functions. One challenge of the protection of dc systems is locating a fault on the dc line. This is especially the case for the systems where the dc link is rather short and ungrounded. The overcurrent-based protection schemes are thus able to localize the faults only to the zones upstream/downstream of the converters. This paper has focused on fault interruption and isolation. The use of sectionalizers to minimize the interruption is the second stage—the reconfiguration stage—of fault management, and it has been investigated in [21]. REFERENCES [1] G. Asplund, “Application of HVDC light to power system enhancement,” presented at the IEEE Power Eng. Soc. Winter Meeting Singapore, 2000. [2] U. Axelsson et al., “The Gotland HVDC light project-experiences from trial and commercial operation,” presented at the CIRED 16th Int. Conf. Exhibition (IEE Conf. Publ. No. 482) Amsterdam, The Netherlands, 2001. [3] M. Baran and N. R. Mahajan, “DC distribution for industrial systems: Opportunities and challenges,” IEEE Trans. Ind. Appl., vol. 39, no. 6, pp. 1596–1601, Nov./Dec. 2003. [4] N. Doerry, H. Robey, J. Amy, and C. Petry, “Powering the future with integrated power system,” Naval Eng. J., vol. 108, p. 12, 1996. [5] P. M. Anderson, Power System Protection. New York: IEEE Press, 1999. [6] B. Zhang and A. Q. Huang et al., “The built-in current sensor and overcurrent protection of the emitter turn-off (ETO) thyristor,” presented at the Ind. Appl. Conf., Salt Lake City, UT, 2003. [7] B. Pauli and G. Mauthe et al., “Development of a high current HVDC circuit breaker with fast fault clearing capability,” IEEE Trans. Power Del., vol. 3, no. 4, pp. 2072–2080, Oct. 1988. [8] Z. Xu and B. Zhang et al., “The emitter turn-off thyristor-based DC circuit breaker,” presented at the IEEE Power Eng. Soc. Winter Meeting, New York, 2002. Mesut E. Baran (S’87–M’88–SM’05) received the Ph.D. degree from the University of California, Berkeley, in 1988. Currently, he is an Associate Professor at North Carolina State University, Raleigh. His research interests include the analysis and control of distribution and transmission systems. Nikhil R. Mahajan (S’01) is currently pursuing the Ph.D. degree in electrical engineering at North Carolina State University, Raleigh. His research focuses on the protection of powerelectronic building blocks. His research interests are in the areas of power system protection and transmission and computer-aided system analysis. Authorized licensed use limited to: IEEE Xplore. Downloaded on April 8, 2009 at 14:03 from IEEE Xplore. Restrictions apply.