Electric Power Systems Research 79 (2009) 984992

Contents lists available at ScienceDirect

Electric Power Systems Research

journal homepage: www.elsevier.com/locate/epsr

Islanding protection of active distribution networks with renewable

distributed generators: A comprehensive survey
S.P. Chowdhury a, , S. Chowdhury a , P.A. Crossley b
a
b

Electrical Engineering Department, University of Cape Town, Cape Town 7701, South Africa
Joule Centre for Energy Research, B24 Pariser Building, University of Manchester, M60 1QD, UK

a r t i c l e

i n f o

Article history:
Received 29 December 2007
Received in revised form 24 July 2008
Accepted 19 December 2008
Available online 3 February 2009
Keywords:
Islanding operation
Protection coordination
Renewable distributed generation
Loss of grid
Digital protection

a b s t r a c t
Anti-islanding protection schemes currently enforce the renewable distributed generators (RDGs) to disconnect immediately and stop generation for grid faults through loss of grid (LOG) protection system.
This greatly reduces the benets of RDG deployment. For preventing disconnection of RDGs during LOG,
several islanding operation, control and protection schemes are being developed. Their main objectives
are to detect LOG and disconnect the RDGs from the utility. This allows the RDGs to operate as power
islands suitable for maintaining uninterruptible power supply to critical loads. A major challenge for the
islanding operation and control schemes is the protection coordination of distribution systems with bidirectional ows of fault current. This is unlike the conventional overcurrent protection for radial systems
with unidirectional ow of fault current. This paper presents a comprehensive survey of various islanding
protection schemes that are being developed, tested and validated through extensive research activities
across the globe. The present trends of research in islanding operation of RDGs are also detailed in this
paper.
2009 Elsevier B.V. All rights reserved.

1. Introduction
With growing power demand and increasing concern about the
use of fossil fuels in conventional power plants, the new paradigm
of distributed generation is gaining greater commercial and technical importance across the globe. Renewable distributed generation
involves the interconnection of small-scale, on-site distributed
energy resources (DERs) with the main power utility at distribution voltage level [1]. DERs mainly constitute non-conventional and
renewable energy sources like solar PV, wind turbines, fuel cells,
small-scale hydro, tidal and wave generators, micro-turbines, combined heat power (CHP) systems, etc. These generation technologies
are being preferred for their high energy efciency (micro-turbine
or fuel cell based CHP systems), low environmental impact (PV,
wind, hydro, etc.) and their applicability as uninterruptible power
suppliers to power quality sensitive loads. Electric energy market reforms and developments in electronics and communication
technology are currently enabling the control of geographically
distributed DERs through advanced supervisory control and data
acquisition (SCADA) [2]. Lasseter and Paigi [3] have discussed how
interconnected DERs can be operated as microgrids both in gridconnected mode and islanded mode.

Corresponding author.
E-mail address: sp.chowdhury@uct.ac.za (S.P. Chowdhury).
0378-7796/$ see front matter 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.epsr.2008.12.012

High degree of penetration of RDGs (more than 20%) as well

as DER placement and DER capacity have considerable impact on
operation, control, protection and reliability of the existing power
utility [35]. These issues must be critically assessed and resolved
before allowing the market participation of RDGs. This is necessary for utilising full distributed generation potential for generation
augmentation, for enhancing power quality and reliability and providing auxiliary services such as active reserve, load-following,
interruptible loads, reactive reserve, restoration, etc. [6].
The area that is critically affected by distributed generator
(DG) penetration is protection coordination of the utility distribution system. Conventional overcurrent protection is designed for
radial distribution systems with unidirectional ow of fault current.
However, connection of DGs into distribution networks convert
the singly fed radial networks to complicated ones with multiple
sources. This changes the ow of fault currents from unidirectional
to bi-directional ones [7]. Besides, the steady state and dynamic
behaviour of the DGs also affect the transmission system operation
[8]. Various impacts of DG connection on existing utility network
protection are listed below:
i)
ii)
iii)
iv)

false tripping of feeders,

nuisance tripping of protective devices,
blinding of protection,
increase or decrease of fault levels with connection and disconnection of DERs,

S.P. Chowdhury et al. / Electric Power Systems Research 79 (2009) 984992

v) unwanted islanding,
vi) prevention of automatic reclosing,
vii) out-of-synchronism reclosure.
Currently, available technical recommendations, viz., G83/1,
G59/1, G75, ETR-113/1, IEEE-1547, CEI 11-20 prescribe that DGs
should be automatically disconnected from the MV and LV utility distribution networks in case of tripping of the circuit breaker
(CB) supplying the feeder connected to the DG. This is known as
the anti-islanding feature in power distribution system protection scheme. This is incorporated as a mandatory feature in the
inverter interfaces for DGs available in the market. As the DGs are
not under direct utility control, use of anti-islanding protection is
justied by the operational requirements of the utilities [9]. Antiislanding protection systems are mainly used to ensure personnel
safety at the grid end and to prevent any out-of-synchronism reclosure.
However, with greater DG penetration, automatic disconnection of DGs for loss of grid (LOG) situation drastically reduces
the expected benets of DGs in (i) maintaining power quality
and reliability, (ii) enhancing system security and (iii) providing
several ancillary services. It also leads to unnecessary loss of DG
power in the event of utility fault [8,9]. Besides, the islanding
detection and anti-islanding protection systems tend to increase
the complexity of protection system. For better utilization of DG
benets, the idea of keeping the DERs connected during system
disturbances and islanding operation and protection of DGs [911]
are being debated upon by researchers across the globe. Various low-cost and efcient digital islanding protection schemes are
being developed, tested and validated through extensive research
activities [12,17]. Fast and efcient microprocessor-based islanding
protection systems are suitable for operation of the active distribution networks both in stand-alone and grid-connected modes.
They can also ensure seamless operation of the inter-tie CBs for
reconnection of the islanded zones without affecting original protection coordination of the utility grid [13]. Improved islanding
protection systems are also being developed for hybrid renewable
energy power systems. These schemes efciently combine the passive (under/over voltage, under/over frequency) and active (Sandia
frequency shift (SFS) and Sandia voltage shift (SVS)) protection
methods [14]. Digital protection schemes are also being designed
and tested for meshed distribution systems with high penetration
of DGs [10].
This paper presents a comprehensive survey of various islanding
protection schemes that are being developed, tested and validated
through extensive research activities across the globe. It also provides a critical assessment of current DG interconnection practice
with the possibility of islanding operation and reclosure.

985

1) It should operate within half a second following the isolation of

the power island, but faster relaying is attractive.
2) The DG units should maintain the systems voltage and frequency within specied limits.
3) It should prevent any out-of-synchronism reclosure.
The LOG detection schemes usually used are as follows:
a) For DGs less than 250 kVA:
For small DGs, LOG is either detected by a reverse power relay
that monitors the power ow in the inter-tie circuit or through
under voltage and under frequency relays. LOG usually causes
severe overloading of the DG unit, causing its output voltage and
frequency to fall. In that case, under voltage and under frequency
relays will operate and automatically disconnect the DG unit from
utility loads.
b) For DGs more than and equal to 250 kVA:
Larger DGs are usually tted with high-speed automatic voltage
regulators. Therefore, their generation may be able to maintain the
voltage and frequency at the load bus within specied limits following LOG. Thus, for such DGs, specialised relaying will be needed
to detect LOG and trip the inter-tie breaker.
2.2. Techniques of LOG detection

2. Loss of grid protection

The direct method for LOG protection is to monitor auxiliary contacts on all CBs on the utility system between its main generation
source and the DG units. When a switching operation produces LOG,
a transfer trip scheme can be used to open the inter-tie CB between
the two systems. The DG units can then be resynchronised to the
utility and reconnected after successful restoration of the utility
supply. But when several CBs are involved in creating the LOG condition, then the transfer trip scheme can only be managed through
an extensive SCADA system and network automation. Reclosure of
the utility supply onto a DG unit can be avoided by using dead circuit pick-up supervision on utility CBs. These would stop the utility
CB breaker from closing until the load-side circuit is de-energised
and would initiate a transfer trip to open the inter-tie CB between
the DG and the utility.
LOG techniques have been classied as active and passive techniques. Active techniques directly interact with the on-going power
system operation, namely (i) reactive power export error detection
and (ii) system fault level monitoring. Passive techniques detect
LOG solely by monitoring the change in power systems behaviour
following such occurrence, namely (iii) under/over voltage and
under/over frequency, (iv) rate of change of frequency (ROCOF),
(v) phase displacement monitoring, (vi) rate of change of generator power output and (vii) comparison of ROCOF (COROCOF). Some
other techniques are (viii) intertripping, (ix) fault thrower and (x)
neutral voltage displacement (NVD).

2.1. Objective and requirements

2.3. Active techniques

The main objective of LOG protection is to detect the condition

where the DGs are left connected to a portion of the utilitys load
network with no main source of utility power. This is commonly
known as LOG phenomenon. This may occur following a system switching operation caused by fault clearance, scheduled and
unscheduled load shedding, maintenance outages and/or equipment failure. The main objective of LOG protection is to detect and
disconnect the inter-tie CB between the DG units and the utility
network after any LOG occurrence. It is necessary to (i) enable an
uncomplicated restoration of utility supply and (ii) avoid any outof-synchronism reconnection of the two systems that may cause
severe damage to the utility grid and the DG units [15]. The typical
requirements for this protection are as follows:

2.3.1. Reactive power export error detection

The reactive power export error detection relay interfaces with
the DG control system to force it to generate a level of reactive power
ow in the inter-tie between the DG and the utility. This level can
only be maintained when the utility generation remains connected.
Relay operation is triggered only when there is an error between
the setting and the actual reactive power being exported for a time
period greater than the set value. In order to avoid any spurious
operation, the time setting is chosen to be greater than the duration
of probable supply uctuations. With operating times from 2 to 5 s,
this is a comparatively slow protection and is frequently used as
backup to other faster protections. Nevertheless, this approach can
detect LOG even when there is no change in the generator loading

986

S.P. Chowdhury et al. / Electric Power Systems Research 79 (2009) 984992

due to the switching operation [15]. Hence, it might be used as a

reliable and effective backup LOG protection technique.
2.3.2. System fault level monitoring
In this scheme, system fault level is monitored by sensing the
short circuit current and reduction in supply voltage when a shunt
inductor is connected across the supply by using point-on-wave
triggered thyristor switches. The ring of the thyristors occurs just
before a current zero. This causes a short pulse of current to ow in
the inductor and a voltage glitch. The tripping decision depends on
the comparison of the measured system fault level with that corresponding to a network fed from the utility generation. The system
can adequately sense the huge difference between the fault levels
of utility generation and that of DG units. It does not jeopardise the
accuracy of operation as it operates by sensing high difference of
fault MVAs. Its tripping depends on the measurement of the value of
the power system source impedance measured close to the inter-tie
[15]. Hence, this scheme provides faster protection with minimum
operating times of about half a cycle.
2.4. Passive techniques

2.4.4. Rate of change of generator power output

With the introduction of digital technology and integrated protection schemes, other techniques are used with more input data for
the decision making. One such technique uses an algorithm based
on the rate of change of power output measured at the DGs terminals. Considering a simplied representation of a DG, its on-site
load and the utility grid as shown in Fig. 1, both the DG and the
utility grid can be modelled by idealised generators of capacity Gg
and Gm and with inertia constants Hg and Hm , respectively.
Any change in system load of Ps produces a change in the
generators loading of Pg as follows:
Pg =

In most circumstances, LOGs result in change in loadings of both

DGs and storage units. Thus, the system voltage, currents and frequency are also changed when LOG occurs. The relays continuously
monitor these system parameters. This is the basis of passive LOG
protection schemes.
2.4.1. Under/over voltage and under/over frequency
Under/over voltage and under/over frequency relaying are effective LOG techniques for small DGs. These are quite direct protection
techniques. They usually provide an acceptable level of protection
for systems with inadequate or no load compensation. But they
might fail to operate if the change of load associated with LOG can
be compensated by the DGs control system, keeping voltage and
frequency more or less stable [15].
2.4.2. Rate of change of frequency
ROCOF relay utilises the generator swing equation to dene
ROCOF as follows:
f
Pf
=
2HG
t

Mention should be made that both ROCOF and phase displacement monitoring schemes have an inherent advantage. In case of
any failure to operate in their rst instance, they trip in immediately subsequent load changes. These relays need only single input
of system voltage waveform and thus require very simple input
circuitry. Hence, they are very much suitable for stand alone relaying.

(1)

where P is the change in power output, f the system frequency, H

the inertia constant of DG system and G is the rated capacity of DG
system.
ROCOF relay monitors the voltage waveform and trips the CB
when the measured ROCOF exceeds a preset value for longer than
a set time period. The relay settings are chosen such that the relay
operates only for uctuations associated with LOG and with the
DG operating independently from the utility, but not for those
uctuations governed by utility time constants. For small- and
medium-sized DG units, a setting of 0.3 Hz/s is found to be optimum
with an operating time from 0.3 to 0.7 s. However, for excessive
changes in frequency, tripping time might possibly be set at less
than four or ve cycles [15]. It suffers from the drawback [16] that
it cannot reliably discriminate frequency changes by LOG as well as
that by other disturbances.
2.4.3. Phase displacement monitoring
The phase displacement monitoring is generically related to the
ROCOF relay operation. It operates when there are phase displacements in the voltage waveform. Such relays are sensitive to direct
changes in DG loading and might cause them to operate in 50 ms
following a LOG resulting in a load change of greater than 5% of the
DG capacity [15].

Ps (Hg Gg )
(Hg Gg + Hm Gm )

(2)

However, if the DG system operates independently from the utility

grid, any load change would directly affect the DG as follows:
Pg = Ps

(3)

An LOG condition usually causes a change in DG loading which

in turn alters the instantaneous DG power output. This method
detects LOG by directly monitoring the changes in the DG output.
Instantaneous DG power output is given by the equation
Pg = va ia + vb ib + vc ic

(4)

where va , vb and vc are the sampled values of phase voltages and

ia , ib and ic are sampled values of the line currents measured at the
DG terminals. With these measurements, the protection algorithm
monitors the changes in power output and integrates these changes
over a dened sample period. Tripping occurs when the integrated
signal exceeds the trip setting. This approach can quickly detect
out-of-synchronism reconnection of the utility supply to a power
island containing DGs [15].
The protection algorithm monitors the changes in power output
and integrates these changes over a dened sample period. Tripping
occurs when the integrated signal exceeds the trip setting, Kg as
shown below.
n=0


(Pg )n > Kg

(5)

n=tx

where n is the sampling instant of Pg and tx is the length of the

sampling window.
This technique is quite safe from malfunctioning during any
extreme load unbalance or loss of phase operation. This is because
of its integrating feature. Any unbalance in the input waveforms
would introduce sinusoidal terms of twice the power frequency and
it might be ltered out by the integrating process. The length of the
sampling window is chosen [15] to give a maximum operating time
of six cycles, and (Pg )n is amplitude limited to give a minimum
tripping time of one cycle. The setting for the relay is chosen such
that it trips whenever the disturbance produces a load change of
1% of the DG rating. In addition to an effective detection of LOG,
this technique would also operate for any subsequent load disturbance should the LOG fail to produce the disturbance required to
trip. Besides, it would trip even if the systems are reconnected while
the systems are out-of-synchronism.

S.P. Chowdhury et al. / Electric Power Systems Research 79 (2009) 984992

987

Fig. 1. Typical network conguration with DG and utility grid.

2.4.5. Comparison of rate of change of frequency

Like ROCOF relays, COROCOF relays [16] assume that LOG would
cause a sudden change in system. But COROCOF is immune to any
such maloperation as suffered by ROCOF as explained in subsection (2.4.2). As explained by Bright [16], it can discriminate between
changes in frequency due to LOG and changes due to widespread
system disturbances by using blocking signal. During any change in
frequency affecting the entire system, the COROCOF sending relays
transmit a blocking signal to all COROCOF DG protection relays to
prevent relay operation. By contrast, LOG would cause a change in
frequency at the DG but would not normally affect the frequency
of the system as a whole. This also assumes that the power island
formed is very small compared to that of the rest of the system. In
case of LOG, no blocking signal would be generated and the COROCOF protection would therefore trip. However, there is a risk if a part
of the system suffers from an LOG while at the same time the rest
of the system suffers a frequency transient, generating a blocking
signal. However, the relay may fail to operate if a LOG phenomena
and a system frequency transient overlap, until the frequency transient in the rest of the system dies out. This may be prevented by
encoding or modulating the blocking signal according to the system frequency measured by the COROCOF sending relay. On the
contrary, a COROCOF DG protection relay would measure the DG
frequency and would operate only if there is a difference between
that and the system frequency indicated by the blocking signal.
Basically, the protection compares the ROCOF at the generator with
that of the rest of the system, hence the name COROCOF. Similar
principles could be applied to vector shift protection, which also
suffers from similar spurious operation.
2.4.6. Intertripping
This technique is quite different from all the other techniques,
since it does not operate on the basis of measurement of electrical parameters. It detects LOG by sensing the contact openings
at the point of disconnection. Then it transmits this signal to all
DG sites supporting the respective island zones as direct acting
trip command. The technique is direct but it needs to transmit
the signal through any of the following dependable mediums: (a)
leased communication channel scheme, (b) radio or microwave
intertripping scheme, (c) powerline carrier intertripping scheme

or (d) hardwired intertripping scheme with bre optic or copper

cables. These are reasonably immune from interference over distances up to 50 km. In order to have a compromise between cost
and reliability, powerline carrier tripping is used for long HV lines
(132 kV) and leased communication channel scheme for MV and
LV lines. Table 1 [12] shows a quick comparison of the aforesaid
mediums.
It is also necessary to check for remote intertrip signals, whether
the communication media is out of the control of the user and is
independent of main supply. This is required to ensure signal continuity at times of disturbance and network failure [12,17]. Assuming
similar order of dependability and security irrespective of generator size, the cost of intertripping remains the same for different DG
sites irrespective of generator ratings. Normally, the smaller DGs
are less likely to have any intertripping from 33 kV utility grid. Nevertheless, the cost concern is mainly focussed on smaller DGs below
2 MW, because the cost impact of intertripping might be more than
50% of the total connection cost.
2.4.7. Fault thrower
This scheme is a further intertripping possibility, conceptually
being similar to powerline carrier scheme. It employs an automatically operated switch to create a short circuit of one line to earth. It is
a reliable, low-cost, acceptable option for intertripping of rural lines
with several DGs. It is used to trip islanded DGs from the source end
and is interlocked with the island initiating CB. It applies a line-toearth fault which is detected by the NVD relays, as discussed in the
following subsection (2.4.8), leading to tripping of all the islanded
DGs. This acts as the only earth reference with the interlocking, preventing substantial current owing from the grid source. The device
reset time should be coordinated with the network protection and
any auto-reclosing facility [12,17].
2.4.8. Neutral voltage displacement
It is not a standard LOG protection but might be used in conjunction with LOG protective devices for mitigating the risk associated
with their non-operation. A specic requirement of Electricity Supply Regulations (ESR) is to ensure system protection against any
network neutral unearthing. The principle of NVD operation is
based on the neutral voltage shift at any earth fault. The NVD relay

Table 1
Comparison of intertripping mediums.
Case

Medium

Range

Dependability

Hardware cost

Media cost

a
b
c
c
d
d

Leased land line

Radio
Microwave
Powerline carrier (PLC)
Fibre-optic cable
Copper cable

Unlimited
10 km line of sight
20 km line of sight
100 km (overhead line)
2050 km without repeater
10 km without repeater

Regional variations; generally good

Possible atmospheric attenuation
Possible atmospheric attenuation
Good on overhead line
Very good
Good depending on screen and segregation

Low
Moderate
Moderate
High
Moderate
Low

Moderate
Low
Unknown
Nil
Very high
High

988

S.P. Chowdhury et al. / Electric Power Systems Research 79 (2009) 984992

is connected in the delta winding of the duly earthed ve-limb

star-delta transformer whose primary star winding is connected
in the connecter line between the utility grid and the DG system.
The transformer primary is referenced to earth at its star-point
through an earth electrode. A major part of the cost of NVD protection comprises the cost of the VT capable of providing the zero
sequence or neutral displacement voltage reference. The NVD relay
does not trip if there is no earth fault. This is irrespective of the DG
being in grid-connected mode or in islanded mode. But it can readily detect any earth fault irrespective of the mode of operation of
DG, though it is more sensitive to the islanded earth fault. The NVD
relay must be properly coordinated through suitable time discrimination to prevent any spurious tripping for earth faults on other
circuits. Although it is not meant for LOG protection, NVD reliably
and directly protects against the unsafe condition where one line of
the islanded system is referenced to earth leading to over voltage
condition in the other two phases [12,17].
2.5. Comparison of different LOG techniques
In the perspective of increasing DG penetration and consequent
conversion of passive distribution networks to active ones, LOG
protection is arguably a big challenge in protection coordination.
Several different techniques are available for protecting a DG unit
against LOG. These include network supervision, active and passive
techniques. Active techniques directly interact with the operation
of the power system, while passive techniques rely on monitoring
the power systems behaviour by measuring system parameters.
Network supervision methods are of direct type offering high level
of protection [12,17]. However, this is a costly affair for most of the
utility networks where SCADA systems do not cover all of the distribution networks. Active techniques are generally very effective
in detecting LOG. But the signicant disadvantage is their requirement of a direct inuence on the power system. Passive techniques
avoid this problem and are generally quite cheap to install. However,
they cannot ensure guaranteed operation under all LOG situations.
The results of several power system studies have presented to
demonstrate the fundamental capabilities and limitations of using
supply voltage, frequency and ROCOF for LOG protection. The performance of the algorithm using the rate of change of power is quite
promising. It has got a signicant advantage of its ability to quickly
detect out-of-synchronism reconnection of the utility supply to a
DG power island. This is considerably an onerous requirement, since
out-of-synchronism reclosure represents the greatest danger to the
operation of power islands.
One more possibility is the national broadcasting of grid frequency for comparing it with the local frequency of the DG sites. The
DG sites around the country will compare their measured local frequencies with the grid frequency broadcast and use the divergence
from the selected reference level to trip the DG units. However,
this method runs the serious risk of widespread common mode
tripping.
3. Islanding protection
The most challenging aspect of designing the electrical power
system with cogeneration was anti-islanding protection [18]. Major
concern of utility grid operation in parallel with DGs is that the
DGs do not operate under the direct control of utility grid. The
most important operational requirement of the utility is avoiding
accidental islanding of any DG site from the utility power source.
According to utility, a power island thus formed may complicate the
orderly reconnection of the power supply network and may present
a potential hazard to the public and utility personnel. Besides, DG
power supply can deviate from the required quality standards. Several guidelines have been introduced [18] in G59 to ensure that

the presence of the small DGs would not detract from the quality of supply to all customers connected to the system. Several
of these guidelines include the need to provide islanding protection.
The principal objective for an islanding protection is to detect
the power island condition to trip the inter-tie breaker between
the power island and the utility grid. Thus, the power island would
not affect the orderly restoration of the utility supply to the rest
of the network. Since the inter-tie breaker is used to connect two
active systems, hence the power island can be reconnected to
the utility after the network supply is established. The tripping
time for the islanding protection should be critically decided such
that the two systems are successfully separated rst before any
out-of-synchronism reconnection attempt by automatic reclosure.
Currently, the maximum separation time has been specied as 0.5 s;
however, best target tripping time is 0.125 s. Islanding protection is
quite complicated, since the CB that causes the LOG could be any
breaker, switch or isolator connecting the main utility grid to the
DG site. Besides, it is unlikely that the status of the CB is supervised
or tted with synchronism checking or live line/dead bus and live
bus/dead line supervision.
The economics of small DG schemes are such that once the
desired level of protection is dened, the protection cost does not
vary with the DGs capacity. Besides, the available utility grade
relays are quite costly for DGs. Thus, for economic viability, there
is obvious necessity for low-cost, high-quality protection packages
for the complete protection of the DG system. The use of a single
microprocessor based relay with integrated protection functions,
including islanding protection, as per the need of the system, is an
ideal solution.
4. Protection scheme for meshed distribution systems
Disconnecting DGs by LOG protection at every fault inception,
even for temporary faults in overhead lines, makes the distribution system unreliable. Viawan et al. [10] have reported, through
case studies, that a protection scheme for meshed distribution systems might be employed to prevent disconnection of DGs during
utility faults. They emphasise that this meshed operation does not
put the DGs or any part of the distribution network in islanding
operation. It also ensures that conventional overcurrent protective devices namely, CBs, relays, reclosures, fuses, etc., do not loose
their properly coordinated functions. Hence, the distribution systems with high penetration of DGs are better operated in loop or
mesh. However, this scheme might lengthen the duration of fault
affecting systems dynamic behaviours, which need to be further
investigated.
The basic scheme connects the DGs to two feeders, which are
operated in a loop by closing a normally open (NO) switch. The DG
is disconnected from the faulted feeder at the fault inception, while
it remains connected to the sound feeder for reliable supply to the
distribution system. The number of CBs needed and their placement
depend on the connection points of the DGs. The protection scheme
performs the following tasks:
(a) It breaks up the loop in order to isolate the DG from the faulted
area and puts the faulted feeder into radial operation.
(b) It clears the fault in the faulted section.
(c) It puts the feeders back to meshed operation once the temporary fault is cleared.
The above functions might be performed by a microprocessorbased high-speed digital line protection relay with the following
featurespilot and non-pilot distance protection, phase and ground
directional overcurrent protection, synchronism check, reclosing,
breaker failure and PLC-based I/O. These relays are available in

S.P. Chowdhury et al. / Electric Power Systems Research 79 (2009) 984992

the market but before application, their cost impact needs to be

critically judged.
Celli et al. [20] have also indicated that revolutionary changes
are needed in the operation and planning of distribution networks
with high penetration of DGs through efcient DG placement and
adoption of meshed network architecture. They performed steady
state and dynamic simulation of a portion of Italian distribution
network to analyse the voltage prole, power losses, short circuit
currents and grid reliability. They have also veried and concluded
that closed-loop networks can easily be implemented. However, for
extracting maximum benets, some modications must be made
in protection coordination by limiting short circuit currents and
applying a new voltage control. The operation of meshed networks
is a valid option to improve networks capability to accommodate DGs, but their development should accompany revolutionary
changes in distribution network operation. Although changing from
radial to meshed networks is advantageous from DG diffusion point
of view, for thick meshes, DG placement must carefully be planned
to get maximum benet.
Meshed networks equalise voltage along feeders and help to
achieve acceptable voltage regulation. In order to have better
voltage regulation, each DG should contribute to regulation and
thus voltage regulator should be modied accordingly. There is
hardly any publication on the studies on protection coordination in
non-radial networks with DGs. However, Brahma and Girgis have
proposed [2123] an adaptive scheme to protect distribution systems with DGs. This scheme can easily be implemented in a network
but requires enormous efforts in modifying a well-consolidated
conventional operating practice.
Brahma and Girgis [23] have concluded the following for
microprocessor-based protection coordination of distribution systems with high penetration of DGs.
a) In case of high DG penetration, coordination between reclosure
and fuse is likely to be disrupted.
b) This problem cannot be resolved with available traditional reclosures.
c) Microprocessor-based reclosures are available in the market to
resolve this problem completely.
d) Protection coordination with microprocessor-based reclosures
is simple.
5. Low-cost microprocessor-based distribution system
islanding protection
Islanding protection is quite complicated because the LOG might
take place by the operation of any CB, switch or isolator connecting
the utility grid to the DGs. Besides, the status of the CB is not normally supervised or tted with synchronism check or live line/dead
bus and live bus/dead line supervision systems. The t-and-forget
approach for connecting DGs to utility distribution system is no
more valid and circumstances call for a policy of integrating DGs
into power system planning and operation through active management of distribution networks applying suitable concepts.
Since the utility grid relays are too expensive to be applied
for DG protection, there is an obvious need for low-cost, highquality protection packages for islanding protection. It is reported
by Redfern et al. that the use of a single microprocessor-based relay
with integrated protection functions is the suitable solution to this
requirement [13]. Moreover, the most direct and effective method
for islanding protection is to take the advantage SCADA system for
monitoring auxiliary contacts of all the circuit breaking devices
present in the system between DG site and utility grid. A transfer
trip signal can be provided to open inter-tie CB [13].
Redfern et al. [13,15] have reported a microprocessor-based
islanding protection technique to detect power island condition of

989

DGs. After detection, it quickly disconnects the DGs from the utility grid by tripping the inter-tie breakers between the utility grid
and the DGs. It does not affect the orderly restoration of the utility
supply to the rest of the network. Since the inter-tie breaker is used
to connect two active systems, the DGs can be reconnected back to
the utility grid after the LOG situation is over.
The algorithm in [13,15] detects islanding by monitoring uctuations in the DG power output caused by disturbances and by
verifying the differences between the DG responses experienced
in grid-connected and in islanded mode. This scheme calculates
instantaneous power from DGs site and then derives the rate
of change of power which is then limited in amplitude by specic insensitivity function to overcome subtransient responses of
the DGs. This clipped signal is integrated over a moving window
of a specic length and tripping is initiated when the absolute
value of the integrated signal exceeds the trip setting. The tripping
time is kept sufcient to allow the two systems to be successfully separated before any automatic reclosing attempts occur. This
avoids the chances of any out-of-synchronism reclosure. Maximum
separation time is specied as 0.5 s. However, new protection algorithms are being thoroughly reviewed with a target tripping time
of 0.125 s.
This scheme has been checked for tripping for both islanding and
local load changes during independent operation. The algorithm
remains stable during large local load changes, while utility grid
remains connected and also during local power system fault conditions. This scheme optimises the use of relay-processing capability
under normal conditions enabling it to be included in a low-cost
microprocessor-based protection scheme.
6. Islanding protection method for hybrid renewable DG
Robitaille et al. [14] have reported the development of an islanding protection scheme of a hybrid power system with wind turbine,
photovoltaic array and fuel cell in Hydrogen Research Institute,
Canada. In the scheme, passive and active protection methods,
e.g., under/over voltage, under/over frequency, SFS and SVS are
judiciously combined together to implement islanding protection.
Simulation results for different critical operating conditions indicate that clearing times can surpass those dened by IEEE 1547
standard.
In this scheme, the under/over voltage and under/over frequency
techniques have been used as a basic protection for the system. The
inverter disconnects from the grid when the voltage or frequency at
the point of common coupling (PCC) crosses the predened thresholds. The PCC voltage amplitude should be within 88110% of the
nominal value and its frequency should stay within 59.560.5 Hz as
per Canadian standard C22.2 No107.1-01.
A combination of SFS and SVS methods are also incorporated in
the scheme to utilise the benets of both the schemes. The advantages of SFS or active frequency drift with positive feedback (AFDPF)
are (a) very narrow non-detection zone (NDZ), (b) high efciency,
(c) easy implementation and (d) low cost (even when inverters
are connected to the same PCC). However, it tends to reduce the
quality of output current waveform which may produce instability
when the DG is connected to a weak grid. Otherwise, it provides
a good compromise between efciency of detection, the waveform
quality and the effect of transient behaviour on the system. The
efciency of SFS method decreases signicantly with high-quality
factor loads. This may be corrected by adding the SVS method. The
main advantage of this method is that its efciency is independent
of load quality factor, thus complementing the SFS method. Though
SVS method also slightly degrades waveform quality of the output
current, it has been veried that the combination of SFS and SVS
methods is very efcient in detecting islanding with a very small
NDZ.

990

S.P. Chowdhury et al. / Electric Power Systems Research 79 (2009) 984992

7. Multi-Agent protection scheme for distribution systems

with DG penetration
Perera and Rajapakse [19] have proposed a multi-agent-based
protection scheme for distribution systems with DGs. The network is divided into several segments. Fault detection and isolation
are performed by installing relay agents at points of interconnection between different network segments. The relay agents
communicate with the neighbouring agents through synchronous
communication networks. This scheme uses a wavelet transform
technique to identify the direction of fault current with respect
to a node in the network. In case of a fault, the assigned relay
agents collaboratively determine the faulted zone by acquiring the
feedback signals from CTs, on the interconnected branches, that
measure the currents leaving the node. Wavelet transform coefcients (WTCs) are calculated after measuring the transient currents
in these branches. A fault is designated to be internal or external
from the sign of the WTCs of the currents measured at all points.
When the sign is same, then fault is designated to be internal, otherwise external. In case of external faults, the fault direction can also
be determined from the fact that WTCs of the currents measured
on the faulted branch have a sign opposite to that of the currents
measured in the other branches.
For a fault on any designated segment, relay agents determine
the direction of the fault current with respect to its location and
communicate this information to adjacent relay agents. On the basis
of this information, the relay agents determine the faulted segment
for issuing correct trip signals to the relevant CBs to isolate the
faulted segment. In this algorithm, a relay agent identies a fault
on the busbar at its location as an internal fault. In such a case,
the fault is immediately cleared by the relay agent tripping all the
CBs connected to the busbar and communicating its decision to the
other relay agents.
Wan et al. [24] have also proposed a similar multi-agent protection coordination scheme and veried its application feasibility.
The scheme uses the substation as one JADE Agent Container consisting of a substation management agent and a number of relay
agents, DG agents and equipment agents. The coordination strategy
is embedded in every relay agent. This helps to coordinate the relay
agents taking into account all pre-fault, during-fault and post-fault
constraints. In the coordination strategy, relay settings and time are
not the only parameters that decide the relay coordination. These
agents communicate amongst themselves and also with substation
management agent, DG agents and equipment agents to obtain a
successful coordination.
The validity and effectiveness of this scheme have been
demonstrated by applying it to an agent-based platformJADE.
Communication simulation shows that successful information
communication between agents can be achieved with this scheme.
A major advantage of this scheme is its ability to self-check and selfcorrect. It can also act rapidly and can provide highly selective fault
regions backup function when the primary protection fails. Effort is
on to improve the performance of the multi-agent system to cope
with protection coordination in a more complex system.
8. Islanding operation
Power system islanding operation has been gradually receiving
more attention from researchers, operators, regulators and other
stakeholders for improving distribution system security and economy against the consequences of blackout. During the periods
when the utility supply system is under stress or in LOG situations, a part of the distribution network might be disconnected
from the utility grid to form a power island, containing both distributed generation and loads. As discussed above, the regulatory
enforcement does not allow islanding operations within a utility

network for maintaining power quality, ensuring safety and avoiding unearthed operation and out-of-synchronism reclosure of the
island onto the main power system. Out-of-synchronism reclosure
is a major concern because it causes large currents, end winding forces and torques in rotating machines leading to equipment
failure. Besides, high currents may trip the protective relays causing system shutdown, maloperation or damage of system loads.
However, regulators and stakeholders would probably be interested to expedite re-structuring of regulations in favour of islanding
operations if the problem of out-of-synchronism reclosure could
be tackled. Best et al. have proposed in reference [25] a feasible
solution by synchronous islanded operation of DG to alleviate any
out-of-synchronism reclosure. The basic concept is the operation of
synchronous island in synchronism with the utility grid even without any electrical connection. Hence, the two systems may easily be
reconnected together without any danger of out-of-synchronism
reclosure. This can be achieved by proper DG control and a suitable communication infrastructure as discussed in the following
sections.
9. Synchronous islanding
The authors of reference [25] have validated their proposal of
synchronous islanded operation with their own phase difference
control algorithm applied in an island with a single-set scenario on
a 50 kVA diesel generator using two different governors. In addition to the standard product variable gain governor of the diesel
generator, they have developed another governor which utilises
supplementary inputs in addition to engine speed. It has been
observed that phase difference can be controlled within acceptable
limits, both in steady state and after load disturbances.
Since disconnection of RDGs from the grid in LOG situations prevents maximum benets being drawn from this resource, operation
of power islands in distribution networks is under consideration of
some utilities as a way of improving the continuity of power supply. With successful implementation, power island operation has
the potential for becoming a commonplace in future.
The basic idea of islanding operation is suggested by Ding et al.
in reference [27]. The power island is kept in synchronism with the
utility grid by using a reference signal transmitted from a secure
part of the network. Since the island is held in synchronism, it
can be reconnected to the main system at any time with minimal transient effect. To its advantage, power island is not limited in
geographical size. Technical advancements in communication techniques and phase measurement units (PMU) along with the advent
of fast digital governing of generation are helpful in the successful operation of power island [2529]. It can sustain supply to the
trapped load within the island, if the generation capacity of DGs
is sufcient. Large load disturbances may cause frequency deviations regardless of the control scheme, leading to its deviation from
synchronisation. Hence, small aggregated loads are more suitable
than any single big load whose capacity is a signicant percentage
of the generation capacity. Therefore, suitable loads may include
small islands in residential and commercial areas or larger islands
of mixed load type, but should exclude small islands feeding large
industrial loads.
9.1. Principle of operation
The principle of operation of synchronous islanding operation is
discussed in the following subsections.
9.1.1. Phase control
Normally, islanding is ideal when the load and generation of the
islanded system are closely matched. Any difference between the

S.P. Chowdhury et al. / Electric Power Systems Research 79 (2009) 984992

991

load and generation of the islanded system causes a frequency transient as the DG increases or reduces its power output [25]. In case of
any load imbalance within the island, a typical generator governor
takes a fraction of a second to settle to the nominal frequency with
a constant lagging phase difference at steady state. In the operation of a synchronous island, a reference voltage waveform signal is
transmitted from a secure part of the localized network. This may
be any point on the utility grid electrically close to the island or
at the main substation of the distribution network. The generator
voltage signal is compared with the reference voltage waveform
to provide a measure of frequency and phase difference between
the two systems. This is used as an input signal to the DG governor to control the frequency and phase of the islanded system. The
steady state lagging phase difference can be made zero by causing a frequency overshoot through phase difference control [25] as
per equal area criterion between negative frequency error and positive frequency error for the removal of phase difference. However,
generating a control signal by just comparing the reference voltage waveform with that of the island may not be enough to keep
the island in synchronous operation. It requires the two voltage
waveforms to be matched correctly in the time domain for having
a zero phase difference at the reconnection point. The delay time in
transmission and calculation of the phase angle variation between
the place of measurement and the reconnection point, as discussed below, are quite important in the operation of synchronous
islands.

10.1. Protection coordination

9.1.2. Delay
The delay for the reference voltage waveform, when transmitted by a short-range FM radio link as in reference [25], is negligibly
small. But this transmission may not be satisfactory for large-scale
implementation of multiple synchronous islands in any distribution
network. Better technique is to use PMU with global positioning system (GPS) time synchronised measurements. Since GPS
measurements are accurate to 1 s with a phase error of 0.018
in a 50 Hz system, the time stamping ensures that the reference
and island signals can be aligned correctly in the time domain.
Typical transmission and processing delays of PMU data through
direct communications are about 100150 ms for bre-optic or
microwave links and 200300 ms for telephone lines.

10.4. Communication failure and security

9.1.3. Phase angle variation

The knowledge of the voltage phase angle variation both in
steady state and transient condition is very important to achieve
successful operation of synchronous island. Phase angle variation
primarily depends on site, load ow and network impedance. It has
been observed for Irelands electricity network [25] that a phase
shift of 30 takes place for the presence of delta-star transformers. For wind farms, there is an abrupt change in the voltage phase
angle at the times of disconnection and reconnection. There is an
8 change when the wind farm is reconnected and the maximum
power output is reached. The variability in voltage phase angle follows the load prole while the wind farm is not connected. The
phase angle may also vary with change in the system structure by
islanding. Since the phase angle varies within a narrow band, it is
suitable for the implementation of synchronous island.

Further investigation is necessary to validate the operation of

multiple power islands using PMU with GPS and supervisory controllers with fast communication infrastructures. Mention may be
made that two-way communications with a supervisory controller
is quite useful to provide the secondary control functions necessary for stable operation of multiple synchronous power islands.
The ability to constantly maintain synchronism for multiple power
islands should also be veried with and without limits of maximum
load disturbances in the islands.

10. Observations
With the development of cost-effective supervisory control and
fast communication infrastructures, synchronous islanded operation is likely to form an integral part of future active distribution
networks after practical demonstrations. Some key observations
about islanding operation, control and protection in relation to
active distribution network management and clean power production are discussed below.

Protection coordination is quite complicated with different

operating conditions under grid-connected mode and power island
mode of operations. The fault current ows are bi-directional
and fault levels are quite different in these two modes of operation. Hence, adaptive protection schemes with high-speed digital
communications would be suitable with proper protection coordination.
10.2. Power quality
Power quality issue should be duly considered in power island
operation. Appropriate measures are to be ensured for voltage and
frequency control, stability, harmonics, unbalanced loads and voltages, earthing and fault clearance.
10.3. Detection of reconnection
It may be difcult to detect the reconnection of the power island
to the utility grid when both are operating in synchronism due
to absence of any feasible transient effect. A reliable detection of
the reconnection to the utility grid would be required for proper
DG control. Since the steady-state phase difference is likely to vary
more during islanded operation than the grid-connected mode, it
can be detected on the basis of the phase difference.

The control system may keep the island in synchronism during

short communication outages by the extrapolation of the reference
signal in place. In case of longer period of communication failures (more than about 2 s), power island needs to be shutdown
for preventing any out-of-synchronism reclosure to the utility grid.
However, redundant communications links might help to alleviate
such situations by using PMU with GPS and supervisory controllers.
Security of information and quality of service should be assured for
using Internet protocol communications in power system control.
10.5. Multiple power islands

11. Conclusion
This paper presents a comprehensive survey of islanding protection of active distribution networks with DG units. The DG paradigm
has created widespread interest in power system planning and
research in recent years amongst energy planners, policy makers,
regulators, generators and researchers. Resolving technical and economic issues related to interconnection of non-conventional and
renewable DERs has been a major thrust of work in this area. The
main challenge of RDG interconnection is the protection coordination of the distribution system with bi-directional ows of fault
current. Protection coordination for such systems is quite different to that for radial systems with unidirectional ow of fault
current. Further challenges are the impacts of steady state and
dynamic behaviour of the DGs on transmission system operation

992

S.P. Chowdhury et al. / Electric Power Systems Research 79 (2009) 984992

and expansion [8]. Initially, when the degree of DG penetration was

low, then anti-islanding feature was imposed worldwide to prevent accidental islanding of DGs. This feature is still in operation
for personnel safety at grid end and for avoiding any untoward outof-synchronism reclosure. Anti-islanding feature forces the DGs to
be disconnected immediately in the event of grid faults by different
LOG-detection schemes.
In recent years, the rapid growth of DGs and RDGs has enhanced
DG penetration across the globe to a very high level. This is gradually leading to reassessment of the anti-islanding protection, as
it greatly prevents utilization of DGs for enhancing power quality
and reliability. Besides, the islanding detection and anti-islanding
protection also increase the complexity of protection systems. In
the perspective of rising consciousness about possible benets of
RDGs, proposals against anti-islanding protections are being established in the form DG islanding protection schemes. The aim of
research is the development of effective microprocessor-based digital schemes for disconnecting the DGs from the utility in case of LOG
situation and allowing them to operate as power islands [12,17].
The synchronous operation of power islands solves the problems of
out-of-synchronism reclosure. Synchronous islands can be reconnected to the utility grid any time once the grid faults are cleared.
Research also aims to make these schemes fast, effective, easy to
implement and as maintenance-free as possible. Apart from technical soundness, research should also aim at developing low-cost
schemes which should be cheap enough to attract the attention of
the DG owners. Current research should also aim to utilise of stateof-the-art communication techniques like GPS, etc., for islanding
protection and control of geographically distributed and remotely
located DG units.
Acknowledgement
The authors are pleased to acknowledge the support provided
by The University of Manchester, EPSRC Supergen Project.
References
[1] N. Jenkins, R. Allan, P.A. Crossley, D. Kirschen, G. Strbac. Embedded Generation,
IEE Power and Energy Series, vol. 31, IEE Press, 2000.
[2] A. Bertani, et al., Electrical energy distribution networks: actual situation and
perspectives for distributed generation, in: Proceedings of the CIRED, 17th International Conference on Electricity Distribution, Barcelona, Spain, May 1215,
2003.
[3] R.H. Lasseter, P. Paigi, Microgrid: a conceptual solution, in: Proceedings of Power
Electronics Specialists Conference (PESC), Aachen, Germany, June 2025, vol. 6,
2004, pp. 42854290.
[4] N. Hadjsaid, J.F. Canard, F. Dumas, Dispersed generation impact on distribution
networks, IEEE Comput. Appl. Power, 1999.
[5] CIGRE, Impact of increasing contribution of dispersed generation on the power
system, Final Report, Study Committee No. 37, September 1998.
[6] T. Moore, Emerging markets for distributed resources, EPRI J. (March/April)
(1998).
[7] L. Kumpulainen, K. Kauhaniemi, Distributed generation and reclosing coordination, in: Proceedings of Nordic Distribution and Asset Management Conference,
NORDAC 2004, Espoo, August 2324, 2004.
[8] J.A. Pecas Lopes, N. Hatziargyriou, J. Mutale, P. Djapic, N. Jenkins, Integrating distributed generation into electric power systems: a review of drivers, challenges
and opportunities, Electric Power Syst. Res. 77 (2007) 11891203.
[9] F.M. Gatta, F. Iliceto, S. Lauria, P. Masato, Modelling and computer simulation of
in dispersed generation in distribution networks. Measures to prevent disconnection during system disturbances, in: Proceedings of IEEE Power Technology
Conference, Bologna, Italy, June 2326, 2003.
[10] F.A. Viawan, D. Karlsson, A. Sannino, J. Daalder, Protection scheme for meshed
distribution systems with high penetration of distributed generation, in: Proceedings of Power System Conference 2006: Advance Metering, Protection,
Control, Communication and Distributed Resources, Clemson, SC, March 1417,
2006, pp. 99104.
[11] T., Thacker, Control of power conversion systems for the intentional islanding of
distributed generation units, Thesis for Master of Science in Electrical Engineering, Virginia Polytechnic Institute and State University, Blacksburg, Virginia,
USA, September 2005.

[12] Econnect, Assessment of Islanded Operation of Distribution Networks and Measures for Protection, ETSU K/EL/00235/REP, 2001.
[13] M.A. Redfern, J.I. Barrett, O. Usta, A new microprocessor based islanding protection algorithm for dispersed storage and generation units, IEEE Trans. Power
Delivery 10 (3) (1995) 12491254.
[14] M. Robitaille, K. Agbossou, M.L. Doumbia, Modeling of an islanding protection method for a hybrid renewable distributed generator, in: Proceedings
of Electrical and Computer Engineering, Canada, May 14, 2005, pp. 1477
1481.
[15] M.A. Redfern, O. Usta, G. Fielding, Protection against loss of utility grid supply
for a dispersed storage and generation unit, IEEE Trans. Power Delivery 8 (3)
(1993) 948954.
[16] C.G. Bright, COROCOF: comparison of rate of change of frequency protection: a
solution to the detection of loss of mains, in: Proceedings of the Developments
in Power System Protection Conference, Publication No. 479, IEE Press, 2001.
[17] Econnect, Islanded Operation of Distribution Networks, DG/CG/00026/00/00,
2005.
[18] L.J. Powell, An industrial view of utility cogeneration protection requirements,
IEEE Trans. Ind. Appl. 24 (1) (1988) 7581.
[19] N. Perera, A.D. Rajapakse, Agent-based protection scheme for distribution networks with distributed generators, in: Proceedings of IEEE Power Engineering
Society General Meeting, Montreal, Canada, June 1822, 2006.
[20] G. Celli, F. Pilo, G. Pisano, V. Allegranza, R. Cicoria, A. Iaria, Meshed vs. radial MV
distribution network in presence of large amount of DG, Proc. IEEE PES Power
Syst. Conf. Exposit. 2 (2004) 709714.
[21] S.M. Brahma, A.A. Girgis, Development of adaptive protection scheme for distribution systems with high penetration of distributed generation, IEEE Trans.
Power Delivery 19 (1) (2004) 5663.
[22] S.M. Brahma, A.A. Girgis, Effect of distributed generation on protective device
coordination in distribution system, in: Proceedings of Large Engineering Systems Conference on Power Engineering LESCOPE01, Halifax, Canada, July 1113,
2001, pp. 115119.
[23] S.M. Brahma, A.A. Girgis, Microprocessor-based reclosing to coordinate fuse and
recloser in a system with high penetration of distributed generation, Proc. IEEE
Power Eng. Soc. Winter Meet. 1 (2002) 453458.
[24] H. Wan, K.K. Li, K.P. Wong, Multi-agent application of substation protection
coordination with distributed generators, Eur. Trans. Electric. Power 16 (5)
(2006) 495506.
[25] R.J. Best, D.J. Morrow, D.J. McGowan, P.A. Crossley, Synchronous islanded operation of a diesel generator, IEEE Trans. Power Syst. 2 (4) (2007) 21702176.
[26] R.A. Walling, N.W. Miller, Distributed generation islanding-Implications on
power system dynamic performance, Proc. IEEE Power Eng. Soc. Summer Meet.
1 (2002) 9296.
[27] X. Ding, P.A. Crossley, D.J. Morrow, Protection and control of networks with
distributed generators capable of operating in islanded mode, Proc. 8th Int.
Conf. Dev. Power Syst. Protect. 2 (2004) 567570.
[28] R.J. Best, D.J. Morrow, P.A. Crossley, Out-of-phase synchronisation of a small
alternator, Proc. IEEE Power Eng. Soc. Gen. Meet. (2428 June 2007).
[29] D.M. Laverty, D.J. Morrow, R.J. Best, P.A. Crossley, Internet based phasor measurement for phase control of synchronous islands, Proc. IEEE Power Eng. Soc.
Gen. Meet., 2008.
S.P. Chowdhury received his B.E.E., M.E.E. and Ph.D. in 1987, 1989 and 1992, respectively. In 1993, he joined E.E. Department of Jadavpur University, Kolkata, India as
Lecturer. He was promoted to Reader and then to Professor in 1998 and 2006, respectively. He visited Brunel University and The University of Manchester, UK, several
times on collaborative research programme. He has published two books and over
125 papers mainly in power systems. He is a fellow of the IET (UK) with C.Eng.,
IE(I) and the IETE(I) and Member of IEEE(USA). He is a member of Membership and
Regions Board (MRB), MRB Finance Committee, Council and the Regional Representative Committee of the IET(UK). He is currently attached to the University of Cape
Town, SA, as an Associate Professor in Electrical Engineering since 2008. (Email:
sp.chowdhury@uct.ac.za)
S. Chowdhury received her B.E.E. and Ph.D. in 1991 and 1998, respectively. She joined
M/S M.N.Dastur & Co. Ltd., as Electrical Engineer and Womens Polytechnic, Kolkata,
India as Lecturer in 1991 and 1998, respectively. She was promoted to Senior Lecturer
in 2006. She visited Brunel University and The University of Manchester, UK, several
times on collaborative research programmes. She has published two books and over
75 papers mainly in power systems. She is a member of the IET (UK) and IE(I) and
Member of IEEE(USA). She is currently attached to the University of Cape Town, SA,
as Senior Research Ofcer in Electrical Engineering. (Email: sunetra69@yahoo.com)
P.A. Crossley is Professor of Electrical Engineering at The University of Manchester.
He graduated with a B.Sc. degree from UMIST, UK, in 1977 and a Ph.D. degree from
the University of Cambridge, UK, in 1983. He had been involved in the design and
application of digital protection relays and systems for more than 25 years, rst
with GEC, then with ALSTOM and UMIST and the Queens University of Belfast,
UK, and currently with The University of Manchester. He is at present the Director of Joule Centre for Energy Research, The University of Manchester, UK. He is an
active member of various CIGRE, IEEE and IET committees on protection. (Email:
p.crossley@manchester.ac.uk)