IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 28, NO.

2, APRIL 2013 779

Fault Detection and Isolation in Low-Voltage

DC-Bus Microgrid System
Jae-Do Park, Member, IEEE, and Jared Candelaria, Student Member, IEEE

AbstractA fault detection and isolation scheme for low-voltage

dc-bus microgrid systems is presented in this paper. Unlike tradi-
tional ac distribution systems, protection has been challenging for
dc systems. The goals of the proposed scheme are to detect the fault
in the bus between devices and to isolate the faulted section so that
the system keeps operating without disabling the entire system. To
achieve these goals, a loop-type dc-bus-based microgrid system,
which has a segment controller between connected components, is
proposed. The segment controller consists of master and slave con-
trollers that monitor currents and control the segment separation,
which include solid-state bidirectional switches and snubber cir-
cuits. The proposed system can detect faults on the bus regardless
of fault current amplitude or the power supplys feeding capacity.
The proposed concepts have been verified by OrCAD/PSpice sim-
ulations and experiments on hardware test bed.
Index TermsDC distribution, fault protection, microgrids,
solid-state switch.
Fig. 1. Conceptual diagram of a dc-bus microgrid system.

I. INTRODUCTION
Microgrid systems can be divided into ac-bus and dc-bus sys-
tems, based on the bus to which the component systems are con-

R ECENTLY, many distributed power systems have been

researched and developed, especially to meet the demand
for high penetration of renewable energy resources, such as
wind turbines and photovoltaic systems. The distributed power
systems have advantages, such as the capacity relief of trans-
mission and distribution, better operational and economical
generation efficiency, improved reliability, eco-friendliness,
and higher power quality [1], [2]. The energy policy of many
governments in the world competitively increases the require-
ment of the penetration of renewable energy resources and
distributed generation (DG). For instance, in the U.S., Cali-
fornia is trying to increase the usage of renewable generation up
to 33% by 2020 [3] and Colorado has set specific requirements
for DG from eligible renewable energy resources [4].
The microgrid system is a small-scale distributed power
system consisting of distributed energy sources and loads,
and it can be readily integrated with the renewable energy
sources [5][7]. Due to the distributed nature of the micro-
grid approach, the connection to the central dispatch can be
removed or minimized so that the power quality to sensitive
loads can be enhanced. Generally, they have two operation
modes: stand-alone (islanded) mode and grid-connected mode.
Manuscript received March 06, 2012; revised October 26, 2012 and De-
cember 22, 2012; accepted January 22, 2013. Date of publication February 12,
2013; date of current version March 21, 2013. Paper no. TPWRD-00238-2012.
The authors are with the Department of Electrical Engineering, University of
Colorado Denver, Denver, CO 80204 USA.
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TPWRD.2013.2243478
nected. The advantage of the ac-bus-based microgrids is that
the existing ac power grid technologies are readily applicable.
However, problems with the ac grid issues, including synchro-
nization, reactive power control, and bus stability, still persist.
DC-bus-based systems can become a feasible solution because
microgrids are small, localized systems where the transmission
loss is negligible. Moreover, they do not have the problems that
ac systems do, and system cost and size can be reduced com-
pared to the typical ac-dc-ac conversion configuration because
dc power is generally used in the power-electronics devices as
a medium. A conceptual diagram of the dc-bus microgrid is
shown in Fig. 1.
While the advantages of dc microgrids are considerable,
the protection of dc distribution systems has posed many
challenges, such as autonomously locating a fault within a
microgrid, breaking a dc arc, dc protective equipment, and
certainly the lack of standards, guidelines, and experience [8].
This paper presents a fault detection and isolation scheme
for a low-voltage dc-bus microgrid system. The goals of the
proposed scheme are to detect the fault in a bus segment be-
tween devices and then to isolate the faulted section so that the
system continues to operate without disabling the entire system.
To accomplish these goals, this paper proposes a loop-type
dc-bus-based microgrid system which has a segment controller
between connected components.
II. LOW-VOLTAGE DC-BUS MICROGRID
Compared to high-voltage dc (HVDC) systems, the
low-voltage dc (LVDC) system is a relatively new concept in

0885-8977/$31.00 2013 IEEE

780
Fig. 2. Fault types in dc systems: (a) line-to-ground fault and (b) line-to-line
fault.
electric power distribution. For small-scale systems, LVDC
microgrids have many advantages over traditional ac distribu-
tion systems. For both ac and dc microgrids, power-electronic
converters are required to connect a variety of sources and
loads to a common bus. Using a dc bus requires fewer stages
of conversion [8][10]. Furthermore, the cables for the ac and
dc power systems are chosen based upon the peak voltage of
the system, and the power delivered by an ac system is based
on the rms values, while the dc power is based on the constant
peak voltage. Hence, the dc system can deliver times that
of an ac system with the same cable. And dc systems do not
suffer from the skin effect. Therefore, the dc system can utilize
the entire cable, thus decreasing losses [9], [11].
Problems arise with dc microgrids when a system needs re-
liable and versatile protection. AC systems have plenty of ex-
perience and standards when it comes to system protection. DC
systems do not have either of these advantages. The switchgear
in dc systems must be very robust in order to handle the dc arc
that is created during the interruption of fault currents. The pro-
tection devices commercially available for low-voltage dc-bus
systems are fuses and circuit breakers (CBs) [8]. Traditional ac
CB mechanisms, which rely on the natural zero crossing of the
ac current to open the circuit, are inadequate to interrupt dc cur-
rents. More important, the fault persists because the operating
time of the CB increases. Allowing a fault current to persist on
a microgrid bus could be catastrophic.
Because the microgrid systems need to be multiterminal,
voltage-source converters (VSCs) must be used to interface
different subsystems to bus. When a fault occurs on the dc side
of a VSC system interfacing the ac source, the insulated-gate
bipolar transistors (IGBTs) lose control and the freewheeling
diodes become a bridge rectifier feeding the fault. The chal-
lenge of protecting VSC systems is that the fault current must
be detected and extinguished very quickly as the converters
fault withstand rating is generally only twice the full-load rating
[12].
III. FAULTS IN DC DISTRIBUTION SYSTEMS
A. Possible Faults
Two types of faults exist in dc systems: 1) line to line and 2)
line to ground, which can be seen in Fig. 2. A line-to-line fault
occurs when a path between the positive and negative line is cre-
ated, short-circuiting them together. A line-to-ground fault oc-
curs when either the positive or negative pole is short-circuited
IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 28, NO. 2, APRIL 2013
to ground. The line-to-ground faults are the most common types
of faults in industrial distribution systems [13].
VSCs may experience internal switch faults that can cause
a line-to-line short-circuit fault. This is a terminal fault for the
device that cannot be cleared; in most cases, the device needs
to be replaced. Hence, dc fuses would be a proper protection
measure for this kind of fault. In ac systems, the ac-side CB
will trip.
B. DC Fault Currents
When a fault occurs in a segment, the line current will split
between load current and fault current
(1)
The magnitude of the fault current depends on the fault loca-
tion and resistance of the fault current path. If the impedance of
fault path is low (e.g., a line-to-ground fault with solid ground),
the current polarity at the receiving end could be reversed, pre-
venting the load from being supported at all. The fault current
from the power source and bus capacitors can be given as fol-
lows:
(2)
where is the line voltage, and are the equivalent re-
sistance and inductance including source, line and ground com-
ponent, and and are the equivalent series resistance
(ESR) and capacitance of bus capacitors, respectively. The time
constant of the dc fault current is quite small because the line
resistance of the dc system is negligible compared to ac power
systems that have high reactance in the line. The bus voltage will
drop or even collapse, depending on the capacity of the power
supply and energy-storage device in the bus, and the grounding
impedance.
C. Current Techniques
The common practice in dc power systems is not to install
any protection on the dc side, and upon fault detection the ac
CBs that link the ac and dc systems are opened [14], [15]. A
handshake method was suggested in [15] to isolate and locate
the faulted segment: however, this method completely de-ener-
gizes the dc system until the fault is removed and the systems
can be re-energized. It works for HVDC and medium-voltage
dc (MVDC) transmission systems where the dc system is a con-
duit between the ac systems and loads. However, this method
can create unnecessary outages in LVDC microgrids where mul-
tiple sources and loads are connected to a common bus. Protec-
tion techniques such as overcurrent [8], [9], [11], current time
derivatives [8], undervoltage and directional protection [9] have
been reported for LVDC distribution systems and dc shipboard
power systems, but the dynamics of voltage and current on a
faulted segment, especially when it is separated has not been
extensively investigated.
Due to the limitations of fuses and traditional ac CBs in
dc systems, a solid-state CB has become a valid option for
dc power system protection. There are several alternatives for
solid-state devices for the CBs, such as gate turn-off thyristors
(GTOs), IGBTs, and insulated-gate commutated thyristors
PARK AND CANDELARIA: FAULT DETECTION AND ISOLATION IN LV DC-BUS MICROGRID SYSTEM 781

(IGCTs). GTOs offer a high-voltage blocking capability and

a low on-state voltage, but suffer from slow switching speeds
[16]. IGBTs are widely used in the low-voltage ( 1200 V)
systems and their advantages include fast interruption time
and high short-circuit current withstanding capability [17].
However, high conduction loss is their disadvantage [18].
IGCTs have the lower conduction losses of a thyristor with the
turnoff capability of a transistor. The IGCT has high voltage
and current ratings as much as IGBTs, and the conduction loss
is relatively low [19]. In order for the microgrid bus to allow
power flow in either direction, the solid-state CB needs to be
bidirectional.
Fuses or molded case CBs (MCCBs) can be utilized on dc
systems [8]. Fuses can be problematic as they will only trip the
faulted line, leaving the unfaulted pole energized. Although this
problem can be solved with MCCBs, the drawback of fuses and
MCCBs is that neither can be controlled autonomously. In the
event of a fault, human intervention is required to re-energize Fig. 3. Conceptual diagram of the proposed protection scheme. Protection con-
trollers in segment B and C are omitted.
the system once the fault has been removed from the system.
Another way to protect the system from excessive fault cur-
rent is to limit the bus current under fault conditions. The fault
current limiters can be used in conjunction with CBs. The ad-
vantages of fault current limiting include ridethrough for tempo-
rary faults, safe operation of switchgear, and system cost down
by switchgear capacity reduction. Several devices have been uti-
lized to limit the fault current, such as superconductors [20], sat-
urated inductors [21], and power-electronics devices [22].

IV. PROPOSED PROTECTION SCHEME

A. Proposed Controller
This paper proposes a novel protection scheme for the dc-bus
microgrid system. Instead of shutting down the whole system or
limiting the bus current, the proposed scheme detects the fault
and separates the faulted section so that the rest of the system
keeps operating. The loop-type dc bus is suggested for the pro-
posed scheme to make the system robust under faulted con-
ditions. It has also been reported that the loop-type bus has a
good system efficiency especially when the distribution line is
not long [23]. The entire loop will be divided into a series of
segments between subsystems. Each segment will consist of a
section of bus (positive and negative lines or positive line and
ground) and a segment controller. The conceptual diagram of
the proposed protection scheme is shown in Fig. 3. The protec-
tion system is shown only in segment A, and controllers on other
segments are omitted.
The proposed protection system consists of one master con-
troller, two slave controllers, and freewheeling branches be-
tween each line and ground. The slave controllers read the cur-
rent at each end of the bus segment connecting two components
and send it to the master controller. They also operate the bidi-
rectional solid-state switches on the bus segment and the free-
wheeling branch according to the commands from the master
controller. In the course of normal operations, the currents mea-
sured at each end of the bus segment should be nearly iden-
tical and the master controller sends commands to put the bus
switches on normal positions.
Fig. 4. Implementation of the proposed protection scheme. Arrows denote the
switching action when a fault is detected.
B. Fault Detection and Isolation
The master controller monitors the difference of two current
readings of slave controllers in a segment
(3)
where and is the line current at each end of the bus
segment. When the difference exceeds the threshold, the con-
troller sends the appropriate commands to slave controllers so
that the faulted segment can be separated from the system. Be-
cause the proposed system uses the differential relaying prin-
ciple monitoring only the relative difference of input and output
current of a segment, it can detect the fault on the bus regardless
of fault current amplitude or power supplys feeding capacity.
Once the faulted segment has been isolated, the bus voltage will
be restored and remainder of the system can continue to operate
on the loop-type bus. Even with multiple faulted segments, the
system can operate partially if the segments from some power
sources to loads are intact. The possibility of the fault around the
device connection point can be minimized if the segment con-
trollers are installed as close to the connection point as possible.
782
Fig. 5. Fault current and operation of the proposed protection scheme. (a) Line-to-ground fault between point A to G. (b) Isolated line-to-ground fault. (c) Line-to-
line fault between point A and B. (d) Isolated line-to-line fault.
The implementation of the proposed protection scheme is
shown in Fig. 4, which shows the configuration of segment A
in Fig. 3. Semiconductor-based bidirectional switches and
diodes are used for segment separation and fault current
freewheeling, respectively. In normal operations, switches
are closed and diodes are open. When a fault occurs, the
master controller detects it using the current information from
the slave controllers and opens switches. Diodes are
conducting at the same time to form a freewheeling path for
fault currents so that the switches can open and the fault
current can be extinguished through resistors. The segment con-
trollers can detect the fault current of the line-to-ground fault
(from point A or B to G) and line-to-line fault (from point A
to B). A turnoff snubber circuit is included for switches to
limit the voltage overshoot due to the line inductance.
The line-to-ground fault and the line-to-line fault are shown
in Fig. 5. It can be seen that the fault current is isolated and
extinguished in the freewheeling loop. The freewheeling path
impedance determines the extinction rate of the fault current,
which can be given as follows:
(4)
where and represent the resistance and inductance in
the freewheeling path, respectively.
When a line-to-ground or a line-to-line fault occurs in the dis-
tribution line, the bus voltage collapse would not allow the load
to ride through if the current is limited from the source because
IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 28, NO. 2, APRIL 2013
of insufficient capacity. This is especially true for the VSC-in-
terfaced microgrid systems. Furthermore, the fault current needs
to be extinguished as quickly as possible even if the system has
sufficient current feeding capacity. Therefore, one of the best
solutions would be to isolate the faulted line as soon as possible
and continue operation with intact bus segments and subsys-
tems. To achieve this, the segment controller needs to be capable
of fast differential current detection and bus switch control. An
automatic reclosing algorithm would be necessary for fault re-
covery and more robust operation [24].
C. Snubber Circuit
Snubber circuits are indispensable to protect the solid-state
CBs from the voltage transient due to the inductance of the
bus cable. It is more so especially in a loop-type bus where
the line inductance exists on both sides of the CB unlike
the point-to-point-type system. Although the fault current
needs to be interrupted as quickly as possible, the high
could make the transient voltage catastrophically high for the
solid-state switches. There are a couple of snubber circuit
topologies to suppress the overvoltage at turn-off due to line
inductance, such as decoupling capacitor, discharge restricted
decoupling capacitor, discharge-charge-type RCD snubber,
and discharge-suppressing-type RCD snubber [25]. It has
been reported that the decoupling capacitor has low losses
but also oscillation issues, and RCD snubbers have higher
losses but no oscillation problem and good for higher current
applications [26]. Since the solid-state CBs do not switch in
PARK AND CANDELARIA: FAULT DETECTION AND ISOLATION IN LV DC-BUS MICROGRID SYSTEM
Fig. 6. Snubber circuits on the source side. (a) Discharge-suppressing-type
RCD snubber. (b) Charge-discharge-type RCD snubber.
high frequency, charge-discharge-type RCD snubber has been
chosen for better voltage suppression performance. Although
the discharge-suppressing-type RCD snubber could eliminate
the separate freewheeling path as can be seen in Fig. 6(a),
it would be difficult to handle bidirectional currents. The
charge-discharge-type RCD snubber is shown in Fig. 6(b).
A case of source-to-load power flow has been considered.
The capacitor in the charge-discharge-type snubber is fully dis-
charged in normal operation when the switch is closed. When
a fault is detected and the switch is trying to open, the diode
is forward biased and the snubber capacitor is charged to
absorb the energy stored in the line inductance . The energy
stored in the line inductance can be given as
(5)
where is the line current. The capacitance of snubber capac-
itor to absorb the energy can be determined as follows:
(6)
(7)
where is the allowable voltage overshoot over . The
snubber resistor forms a discharging path when the switch
is closed. Once the fault has been detected and separated, the
system can check if the fault is cleared by reclosing test. When
the switch is closed, most of the energy stored in the capacitor
(6) should be discharged before the switch is re-opened in case
the fault persists. Hence, the resistance can be determined by
(8)
where is the percentage of stored energy to be discharged in
time before the next switch opening.
D. Grounding
Grounding is one of the most important factors in the power
system safety and protection. For a dc distribution system, the
advantages of the grounding include predictable operating con-
ditions, minimum voltage stress for the system components, and
easier fault detection [27]. The ground fault current can also
be limited by using the resistance grounding. Since the typical
power-electronics converters connected in the LVDC systems
cannot feed large fault currents, it would be beneficial to reduce
783
the fault current to an appropriate level for detection and ex-
tinction. However, some protective devices are still needed even
with the resistance grounding schemes, because the fault current
cannot be sustained.
The dc system grounding options are: 1) solid grounding; 2)
low-resistance grounding; 3) high-resistance grounding; and 4)
no grounding [13]. Although ungrounded systems are used in
some applications to avoid the effect of low-resistance pole-to-
ground fault and stray current, ungrounded systems are sensitive
to changes in the grounding plane and can be dangerous espe-
cially under abnormal fault conditions [28]. International stan-
dard IEC 60364 defines three grounding systems. Low-voltage
power systems ( 1 kVac or 1.5 kVdc) using the two-letter codes
TN, TT, and IT depending on how the source and conductive
body of electric devices are connected to earth [29]. The TN
system ensures that distributed grounding and fault detection is
straightforward but faults at a higher voltage level can increase
the touch voltage. The TT system is easy to install and faults
do not migrate to others but high-voltage stress is possible. The
IT system has very low or even zero fault current and can op-
erate with a single fault, but insulation monitoring is required
and high fault current is possible with another ground fault [30].
It is a common practice to ground power subsystems at a
single point and as close to the source as possible [27], like in a
TN system. Multiple ground points in a TT system could form
unnecessary circulating current paths. Possible grounding point
for a dc system can be either one of the poles or the midpoint of
the bus, and it has been reported that the balanced dc-side mid-
point grounding significantly reduces circulating current com-
pared to the ac-side neutral-grounded system [27]. Furthermore,
it is safer because of lower ground potential compared to the di-
rect pole grounding, and the ground impedance at the midpoint
can limit the fault current under a safe level. The ground resistor
can be used to detect the ground fault as well, although it is not
able to identify the location of the fault because of the single
ground point practices.
V. SIMULATION RESULT
A computer simulation has been performed for a microgrid
system that consists of three typical energy devices: a source, a
load, and energy storage. They are connected as shown in Fig. 7.
Stiff dc power sources are assumed so that a constant fault cur-
rent is fed by the sources without a voltage drop. A 240 V bipolar
dc bus with 200 m bus cable segments [31] and a fault at the
middle (100 m) of the bus is simulated. Simulation parameters
can be found in Table I.
A positive line-to-ground fault in the middle of the bus seg-
ment A is simulated at 1 ms. Fig. 8 shows the source- and
load-side current of a line-to-ground fault with and without pro-
tection. It can be seen that the current from source increased to
180 A after 0.5 msec. The fault current magnitude depends on
the impedance of the fault path. The currents at each end of the
segment which had been identical before fault show clear dif-
ference after fault. Line-to-line fault current will be higher be-
cause there is no resistance to limit it. Therefore, fast detection
and isolation are critical.
It has been assumed in the simulation that the segment con-
trollers can detect it and open/close solid-state CBs in 250 s.
784
Fig. 7. Simulation circuit for the line-to-ground fault in the three-node microgrid system that contains two sources and a load. A fault is simulated in the segment
between source 1 and load.
TABLE I
SIMULATION PARAMETERS
Considering the speed of current microcontrollers and switching
devices, fast interruption in this speed range is feasible. Fig. 8
also shows that the fault currents are extinguished when the
faulted segment is separated.
Fig. 9 shows the voltage across the load and the current in
the freewheeling path. The ground fault pulls the positive pole
voltage to zero, and the bipolar dc bus will experience a voltage
offset on the faulted pole. However, the load voltage is quickly
restored after the faulted segment has been separated.
Fig. 10 shows the voltage transient across the solid-state CB
without and with the snubber circuits, respectively. It also
IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 28, NO. 2, APRIL 2013
Fig. 8. Simulation: Source-side current (top) and load-side current (bottom)
for a line-to-ground fault with and without protection.
shows the fault extinction in the freewheeling path. The voltage
transient at turnoff due to the line inductance and high can
be very high and it can easily damage the solid-state switch. It
can also be seen that the voltage transient is suppressed by a
snubber circuit at a tolerable level.
The protective devices in bus segments B and C have been
omitted in this simulation because the proposed protective
scheme that detects the difference in net current flow will not
be triggered if there is no fault in the bus. This can be verified
PARK AND CANDELARIA: FAULT DETECTION AND ISOLATION IN LV DC-BUS MICROGRID SYSTEM
Fig. 9. Simulation: Load voltage with and without protection (top) and current
in a freewheeling path in the proposed scheme (bottom).
Fig. 10. Simulation: Voltage across switch at turnoff without (top) and
with (bottom) the snubber.
by the currents flowing into and out of the intact bus segment
C shown in Fig. 11. Even with the transient caused by the fault
on the bus segment A, the incoming and outgoing currents
of segment C are identical. The proposed scheme that detects
the current difference is robust to common-mode noise and
transient due to a fault.
VI. EXPERIMENTAL VALIDATION
A lab-scale experiment setup in about 6:1 scale from the sim-
ulation circuit has been built to validate the feasibility of the pro-
posed protection scheme with actual hardware. The setup has a
source-load-storage structure of Fig. 3, but only with positive
pole and ground to simulate a line-to-ground fault. Two chan-
nels of 30-V 3-A power supply Instek GPC-3030D have been
utilized for 20-V source and energy storage, and 12.5- resis-
tive load is used. Each source supplies about 1 A of load current
in normal operation. Three 15- H inductors and six 220-pF ca-
pacitors are installed in the positive line to simulate the line in-
ductance and stray capacitance, respectively. But no resistance
has been added in the bus due to the low current capacity of the
power supply. Two IXYS 600 V IGBT IXGN60N60C2D1 mod-
ules have been used to make a bidirectional solid-state CB and
785
Fig. 11. Simulation: Voltage of bus segments B and C (top) and currents of
segment C and (bottom). Intact segment C currents (incoming) and
(outgoing) are identical.
Fig. 12. Schematic of the experimental setup.
the same IGBT has been utilized to form the ground fault path.
IGCTs can be considered for lower conduction losses. The free-
wheeling path has an IR diode T40HF60 and a 2- freewheeling
resistor. Using a diode in the freewheeling branch simplifies the
control circuits and it can be used for the line-to-ground fault
protection. The RCD snubber circuit is attached to the IGBT
with a 10- F capacitor, a 12- resistor, and a diode T40HF60.
The circuit parameters have not been scaled down because of
the inductance in the power supplies. The schematic of the ex-
perimental test bed is shown in Fig. 12.
The differential protection algorithm described in
Section IV-B is implemented in TIs TMS320F28335 dig-
ital signal processing (DSP) microcontroller. As shown in
Fig. 12, the controller monitors the source and load-side current
using Tektronics A622 current probes through analog-to-digital
channels of DSP, and controls the IGBT CBs. A threshold of
20% has been used for (3) in the experiments, but the threshold
level may vary depending on bus conditions to increase sensi-
tivity or to avoid malfunction. The controller sends out a digital
output to the gate driver board, which generates isolated gating
signal using optocoupler HCPL-3120. In this experiment, the
analog-digital conversion of the current measurement is done
in the 100- s interrupt service routine. Although the controller
takes the analog signal directly from the current sensors because
of the small size of the test bed, digital communication between
the master and slave controller should be considered for larger
786
Fig. 13. Experimental test bed and DSP controller.
Fig. 14. Experiment: Source-side current (top) and load-side current
(bottom) with and without protection.
scale systems. Distributed control without the master controller
can be possible if the two slave controllers are communicating
with each other. The experimental test bed is shown in Fig. 13.
A line-to-ground fault has been experimentally tested. In
normal steady-state condition, the source power supply supplies
20 W of power (20 V, 1 A) to the load and the power-supply
simulating energy storage monitors the bus voltage. When the
ground fault IGBT is turned on, two power-supply channels
feed the fault current, and the segment controller measures
the current shown in Fig. 14. The difference is quite large
because the direction of the load-side current is reversed unlike
the simulation due to the low fault impedance. The controller
readily detects the difference of the bus currents and isolates
the faulted segment immediately. It turned out that the power
supply can feed about 40-A peak transient fault current for a
couple hundred microseconds before each channel settles down
to 3 A of the power supplys rated current. It should be noted
that this power-supply characteristic makes the waveform of
experimental fault current different from that of the simulation
in Fig. 8.
Fig. 14 shows that the fault current is extinguished to zero
as simulated. The fault current extinction time, which depends
on the freewheeling path impedance, was about 100 s in this
IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 28, NO. 2, APRIL 2013
Fig. 15. Experiment: Load voltage with and without protection (top) and cur-
rent in the freewheeling path in the proposed scheme (bottom).
Fig. 16. Experiment: IGBT collector-emitter voltage without (top) and
with (bottom) the snubber circuit.
experiment. Without protection, 6-A fault current flows from
two power-supply channels after transient. Fig. 15 shows that
the load voltage is restored to a normal level quickly after the
faulted segment is separated; otherwise, it collapses to a low
voltage without the protection system. Fig. 16 shows that the
snubber circuit effectively suppresses the voltage overshoot and
oscillation across the IGBT.
VII. CONCLUSION
This paper has presented a fault detection and isolation
scheme for the low-voltage dc-bus microgrid system. The
proposed protection scheme consists of segment controllers
capable of detecting abnormal fault current in the bus and
separating the faulted segment to avoid the entire system
shutdown. A loop-type dc-bus-based microgrid system with
segment controllers between connected components and the
freewheeling branch has been proposed. The proposed pro-
tection concepts have been validated by computer simulations
and experiments. A prototype system with a segment controller
that consists of master and slave controllers has been tested
on an actual hardware test bed and it has shown successful
fault detection and isolation capability. The proposed scheme
can be applied to dc power systems, such as Green Buildings,
PARK AND CANDELARIA: FAULT DETECTION AND ISOLATION IN LV DC-BUS MICROGRID SYSTEM
with sustainable energy resources and data centers with a
server array. Challenges, such as a reduction of conduction
loss in the solid-state CBs and fault ridethrough capability, and
fault-location techniques need to be investigated.
REFERENCES
[1] R. Dugan and T. McDermott, Distributed generation, IEEE Ind.
Appl. Mag., vol. 8, no. 2, pp. 1925, Mar./Apr. 2002.
[2] F. Blaabjerg, R. Teodorescu, M. Liserre, and A. Timbus, Overview
of control and grid synchronization for distributed power generation
systems, IEEE Trans. Ind. Electron., vol. 53, no. 5, pp. 13981407,
Oct. 2006.
[3] U.S. Department of Energy, (2009). States With Renewable Portfolio
Standards. [Online]. Available: http://www..energy.ca.gov/renewables
[4] U.S. Department of Energy Energy Efficiency and Renewable Energy
News, (2010). Colorado Boosts its Renewable Energy Requirement to
30% by 2020. [Online]. Available: http://apps1.eere.energy.gov/news/
news\_detail.cfm/news\_id=15878
[5] R. Lasseter and P. Paigi, Microgrid: A conceptual solution, in
Proc. 35th Annu. IEEE Power Electron. Specialists Conf., 2004, pp.
42854290.
[6] H. Nikkhajoei and R. Lasseter, Distributed generation interface to
the CERTS microgrid, IEEE Trans. Power Del., vol. 24, no. 3, pp.
15981608, Jul. 2009.
[7] F. Katiraei, R. Iravani, N. Hatziargyriou, and A. Dimeas, Microgrids
management, IEEE Power Energy Mag., vol. 6, no. 3, pp. 5465, May/
Jun. 2008.
[8] D. Salomonsson, L. Soder, and A. Sannino, Protection of low-voltage
DC microgrids, IEEE Trans. Power Del., vol. 24, no. 3, pp.
10451053, Jul. 2009.
[9] M. Saeedifard, M. Graovac, R. Dias, and R. Iravani, DC power sys-
tems: Challenges and opportunities, in Proc. IEEE Power Energy Soc.
Gen. Meeting, July 2010, pp. 17.
[10] R. Cuzner and G. Venkataramanan, The status of DC micro-grid pro-
tection, in Proc. IEEE Ind. Appl. Soc. Annu. Meeting, Oct. 2008, pp.
18.
[11] P. Salonen, P. Nuutinen, P. Peltoniemi, and J. Partanen, LVDC dis-
tribution system protection: Solutions, implementation and measure-
ments, in Proc. 13th Eur. Conf. Power Electron. Appl., 2009, pp. 110.
[12] J. Candelaria and J.-D. Park, VSC-HVDC system protection: A re-
view of current methods, in Proc. IEEE/Power Energy Soc. Power
Syst. Conf. Expo., Mar. 2011, pp. 17.
[13] J. Das and R. Osman, Grounding of AC and DC low-voltage and
medium-voltage drive systems, IEEE Trans. Ind. Appl., vol. 34, no.
1, pp. 205216, Jan./Feb. 1998.
[14] P. Cairoli, R. Dougal, U. Ghisla, and I. Kondratiev, Power sequencing
approach to fault isolation in dc systems: Influence of system parame-
ters, in Proc. IEEE Energy Convers. Congr. Expo., 2010, pp. 7278.
[15] L. Tang and B. Ooi, Locating and isolating DC faults in multi-terminal
DC systems, IEEE Trans. Power Del., vol. 22, no. 3, pp. 18771884,
Jul. 2007.
[16] H. Iwamoto, K. Satoh, M. Yamamoto, and A. Kawakami, High-power
semiconductor device: A symmetric gate commutated turn-off
thyristor, Proc. Inst. Elect. Eng., Elect. Power Appl., vol. 148, no. 4,
pp. 363368, Jul. 2001.
[17] Y. Ito, Y. Zhongqing, and H. Akagi, DC micro-grid based distribu-
tion power generation system, presented at the 4th Int. Conf. Power
Electron. Motion Control, Xian, China, Aug. 2004.
[18] R. Schmerda, S. Krstic, E. Wellner, and A. Bendre, IGCTs vs. IGBTs
for circuit breakers in advanced ship electrical systems, in Proc. IEEE
Electric Ship Technol. Symp., Apr. 2009, pp. 400405.
[19] P. Steimer, O. Apeldoorn, E. Carroll, and A. Nagel, IGCT technology
baseline and future opportunities, in Proc. IEEE/Power Energy Soc.
Transm. Distrib. Conf. Expo., 2001, vol. 2, pp. 11821187.
787
[20] F. Luo, J. Chen, X. Lin, Y. Kang, and S. Duan, A novel solid state fault
current limiter for DC power distribution network, in Proc. IEEE 23rd
Annu. Appl. Power Electron. Conf. Expo., Feb. 2008, pp. 12841289.
[21] W. Fei, Y. Zhang, and Z. Lu, Novel bridge-type FCL based on self-
turnoff devices for three-phase power systems, IEEE Trans. Power
Del., vol. 23, no. 4, pp. 20682078, Oct. 2008.
[22] U. Ghisla, I. Kondratiev, and R. Dougal, Branch circuit protection for
DC systems, in Proc. IEEE Elect. Ship Technol. Symp., Apr. 2011, pp.
234239.
[23] M. Saisho, T. Ise, and K. Tsuji, Configuration of DC loop type
quality control center, in Proc. Power Convers. Conf., 2002, vol. 2,
pp. 434439.
[24] J. Machowski, J. Bialek, and J. Bumby, Power System Dynamics: Sta-
bility and Control. Hoboken, NJ: Wiley, 2011.
[25] N. Mohan, T. Undeland, and W. Robbins, Power Electronics Con-
verters, Applications, Design, 3rd ed. Hoboken, NJ: Wiley, 2003.
[26] Y. Zhang, S. Sobhani, and R. Chokhawala, Snubber considerations for
IGBT applications, 2002. [Online]. Available: www.irf.com/technical-
info/designtp/tpap-5.pdf
[27] B. Jacobson and J. Walker, Grounding considerations for dc and
mixed dc and ac power systems, Naval Eng. J., vol. 119, no. 2, pp.
4962, October 2007.
[28] D. Paul, DC traction power system grounding, IEEE Trans. Ind.
Appl., vol. 38, no. 3, pp. 818824, May/Jun. 2002.
[29] IEC 60364-1 Low-Voltage Electrical InstallationsPart 1: Funda-
mental Principles, Assessment of General Characteristics, Definitions,
IEC 60364-1, 2005.
[30] R. Kamel, A. Chaouaski, and K. Nagasaka, Design and testing of three
earthing systems for micro-grid protection during the islanding mode,
Smart Grid Renew. Energy, vol. 1, pp. 132142, Sep. 2010.
[31] Transmission Line Reference Book 345 kV and Above, Elect. Power
Res. Inst., 1987.
Jae-Do Park (M07) received the Ph.D. degree in
electrical engineering from the Pennsylvania State
University, University Park, in 2007.
Currently, he is an Assistant Professor of Elec-
trical Engineering at the University of Colorado
Denver, Denver, CO, USA. Prior to his arrival at
the University of Colorado Denver, he was Manager
of Software and Controls for Pentadyne Power
Corporation, CA, where he took charge of control
algorithm design and software development for the
high-speed flywheel energy-storage system. He is
interested in various energy and power system research and education topics,
including electric machines and drives, energy-storage and harvesting systems,
renewable energy sources, grid-interactive distributed generation, microturbine
control, and microgrid systems.
Jared Candelaria (S10) is currently pursuing the
M.S.E.E. degree in power options at the University
of Colorado Denver, Denver, CO, USA.
He is an Electrical Engineer at NEI Electric Power
Engineering, Arvada, CO. His duties at NEI include
arc flash, interconnection of distributed generation,
supervisory control and data acquisition, substation
design, industrial design, and protective relaying.
His research interests include power systems appli-
cations, power system protection, and HVDC.
Mr. Candelaria is a registered E.I.T. in the state of
Colorado.