The document discusses methodology for selecting, setting, and analyzing anti-islanding protection for distributed generation systems. It explores various anti-islanding detection devices, such as rate of change of frequency (ROCOF) relays and vector surge relays. A literature review covers rationale for anti-islanding protection and different remote and local islanding detection techniques. The project involves developing a MATLAB/Simulink model of a utility network to simulate islanding conditions and evaluate performance of protection devices to assist in selecting appropriate devices and settings.

1. Methodology of Selection, Setting and Analysis of Anti-Islanding
Protection For Distribution Generation System
Kenny SAMAROO
Submitted in partial fulfillment of the requirements for Bachelor of
Engineering (Electrical Engineering)
Electrical Engineering
Faculty of Technology
University of Guyana
August 24, 2012

2. Methodology of Selection, Setting and Analysis of Anti-Islanding
Protection For Distribution Generation System
By
Kenny O. Samaroo
Under the guidance of
Dhanraj Bachai
Department of Electrical Engineering
University Of Guyana
August, 2012
© Kenny O. Samaroo 2012
2|Page

3. ABSTRACT
With the daily increasing demand for power, and need for alternative power generation
technologies, such as, fuel cell, wind & water turbine and photovoltaic systems, customer
demands for better power quality and reliability are forcing the power companies to move
towards distributed generations (DG).
Islanding occurs when a portion of the distribution system becomes electrically isolated from the
remainder of the power system yet continues to be energized by distribution system. It is
important when using DG in an interconnected system that the power distributed system is
capable of detecting an unintentional islanding condition.
Current IEEE interconnection standards (IEEE 1547) mandate that control and protection
measures should be in place to lessen the probability of an unintentional island, and to minimize
the duration of an islanding condition, if one should occur. Typically, a distributed generator
should be disconnected within 100 to 300 ms after loss of main supply [1]. To achieve this each
distributed generator must be equipped with an islanding detection device or anti islanding
devices, such as, vector shift relay and ROCOF relay.
This project seeks to explore the various methods of selecting, setting and analysis of anti-
islanding protection devices (relays) for distribution generation system.
3|Page

4. Table of Contents
ABSTRACT.................................................................................................................................................. 3
List Of Tables ............................................................................................................................................... 5
List Of Figures .............................................................................................................................................. 6
ACKNOWLEDGEMENT ............................................................................................................................ 9
INTRODUCTION ...................................................................................................................................... 10
Background ............................................................................................................................................. 10
Statement Of The Problem...................................................................................................................... 12
SCOPE OF WORK ..................................................................................................................................... 14
Overview ................................................................................................................................................. 14
Literature Review.................................................................................................................................... 15
Rationale for anti-islanding protection: .............................................................................................. 15
Remote Islanding Detection Techniques ............................................................................................ 16
Local Detection Techniques................................................................................................................ 17
METHODS/DESIGN APPROACH ........................................................................................................... 26
Detection of Islanded Power Systems ..................................................................................................... 26
Network Studied ..................................................................................................................................... 27
Simulation model .................................................................................................................................... 30
Conditions for Islanding ......................................................................................................................... 31
Model Description .................................................................................................................................. 33
SIMULATION RESULTS ......................................................................................................................... 46
Normal Conditions .................................................................................................................................. 46
Islanded Condition .................................................................................................................................. 52
Scenario 1: Formation of a Major Island (Loss of Grid) ................................................................... 52
Scenario 2: Formation of a Minor Island. .......................................................................................... 62
CONCLUSION ........................................................................................................................................... 70
RECOMMENATION ................................................................................................................................. 71
BIBLIOGRAPHY ....................................................................................................................................... 72
4|Page

5. List Of Tables
Tables
1. Steady state parameters for the Synchronous Machines used in the model, under normal
operating conditions.
2. Combine results for Relay Protection Blocks 1 and 2 for a major islanded condition.
3. Results obtained for Relay Protection Block 2 for a minor islanded condition.
5|Page

6. List Of Figures
Figures
1. Power system with centralized generation
2. Decentralized power system with DG interconnected
3. Utility Network before and after islanding has occurred.
4. Islanding detection techniques
5. Equivalent Circuit of Synchronous Generator equipped with ROCOF Relay operating
parallel with Utility [4,6]
6. Equivalent circuit of Synchronous Generator equipped with Vector Surge Relay
operating parallel with Utility [4,6]
7. Internal and terminal voltage phasors (a) before opening with CB (b) after opening with
CB.
8. Voltage Vector Surge
9. One line diagram for the Versailles/Lenora portion of GPL’s DIS.
10. Modified Equivalent One Line Network Diagram
11. Matlab/Simulink model of Versailles and Lenora.
12. Distinction between Major and Minor Island and Conditions for Islanding in the Network
Studied
13. Simulink model of a Synchronous Machine
14. Simulink model of a three phase transformer and its equivalent circuit.
15. Simulink model of a three phase source.
16. Simulink model of a three phase breaker.
17. Simulink model of a three phase parallel RLC load.
18. Simulink model of a root mean square (rms) calculation block.
19. Simulink model of a three phase voltage-current measurement block.
20. Simulink model of display block.
21. Simulink model of an oscilloscope (scope).
22. Simulink model for the protection sub-system.
23. Relays found inside the protection sub-system block
6|Page

7. 24. (a) Simulink model of Under/Over Current Relay model.
(b) Simulink model of Under/Over Voltage Relay model.
(c) Simulink model of Under/Over Frequency Relay model.
(d) Simulink model of The Rate of Change of Frequency Relay (ROCOF)
(e) Simulink model of a Vector Shift Relay.
25. (a) Simulation result of Synchronous Machine One (SM1).
(b) Simulation result of Synchronous Machine Two (SM2).
(c) Simulation result of Synchronous Machine Three (SM3).
26. (a) Simulation result for 3 phase voltages and currents at bus 1 and bus 2.
(b) Expanded view of the 3 phase voltages and currents at bus 1 and bus 2.
27. (a) Simulation result for the 3 phase rms voltages at bus 1&2.
(b) Simulation result for the 3 phase rms currents at bus 1&2.
28. Simulation result of the rate of change of frequency and frequency for bus 1 and bus 2.
29. (a) Simulation result for Synchronous Machine One (SM1) for a major islanded
condition.
(b) Simulation result for Synchronous Machine One (SM2) for major islanded condition.
(c) Simulation result for Synchronous Machine One (SM3) for major islanded condition.
30. (a) Simulation result for the 3 phase rms currents at bus 1&2 for major islanded
condition.
(b) Simulation result for the 3 phase rms voltages at bus 1&2 for a major islanded
condition.
31. Simulation result for the 3 phase voltages and current at bus 1&2 for a major islanded
condition.
7|Page

8. 32. Simulation result of the rate of change of frequency and frequency at bus 1 and bus 2
during a major islanded condition.
33. (a) Results obtained from the protection block at bus 1 for a major islanded condition.
(b) Results obtained from the protection block at bus 2 for a major islanded condition.
34. (a) Graph showing comparison of the relays detection time at bus 1.
(b) Graph showing comparison of the relays detection time at bus 2.
35. (a) Simulation result for Synchronous Machine One (SM1) for minor islanded condition.
(b) Simulation result for Synchronous Machine One (SM2) for minor islanded condition.
(c) Simulation result for Synchronous Machine One (SM3) for minor islanded
condition.
36. (a) Simulation result for the 3 phase rms currents at bus 1&2 for minor islanded
condition.
(b) Simulation result for the 3 phase rms voltages at bus 1&2 for a minor islanded
condition.
37. Simulation result of the rate of change of frequency and frequency at bus 1 and bus 2
during a major islanded condition.
38. (a) Results obtained from the protection block at bus 1 for a minor islanded condition.
(b) Results obtained from the protection block at bus 2 for a minor islanded condition.
39. Graph showing comparison of the relays detection time at bus 2.
8|Page

9. ACKNOWLEDGEMENT
I would like to thank the University of Guyana’s Faculty of Technology which provided me the
opportunity to conduct this study. In particular, my supervisor, Dhanraj Bachai, whose
knowledge and guidance played a key role in the success of this work.
I would also like to thank Mr Blackman who provided me with the relevant information needed
to help make this project a success. Also I would like to thank all my class mates for all the
thoughtful and mind stimulating discussions we had, which prompted us to think beyond the
obvious.
Finally I cannot end without thanking my family and more so my wife ‘Priea Samaroo’, on
whose encouragement, support, and advice, I have relied on throughout my studies.
9|Page

10. INTRODUCTION
Background
Electric power industries were traditionally designed with the power distribution system
assuming the primary substation being the sole source of power generation (as shown in Figure
1).
Figure 1: Power system with centralized generation.
With the introduction of Distributed Generation (DG) this assumption changes, that is, power
source/s (DG) are placed within the power distribution system at points where support for active
and reactive power is required after a load flow study is carried out (as shown in Figure 2).
Figure 2: Decentralized power system with DG interconnected.
10 | P a g e

11. Generating power on-site, rather than centrally, reduces cost of transmission, complexity, and
inefficiencies associated with transmission and distribution.
Recently there has been significant increase in the utilization of interconnected DG. The
increasing incursion of DG was driven by improving cost and performance of both old-line and
new technologies, and by customers and third parties seeking to reduce costs, increase local
control of the energy resource, and increasing awareness of the important role of power system
reliability [1].
Distribution generation generally applies to relatively small generating units at or near consumer
site/s to meet specific consumer needs, to support economic operation of the existing distribution
grid, or both. Reliability of service and power quality is enhanced by the proximity to the
consumer and efficiency is often increased.
While central power systems remain crucial to the local utility, their flexibility is limited. Large
power generation facilities are very expensive and require immense transmission and distribution
network to transmit the power. DG compliments central power by providing a relatively low
capital cost in response to incremental increase in power demand while avoiding transmission
and distribution capacity upgrades by placing power source/s within the already existing
grid/network where it is most needed and by having flexibility to send power back into the grid
when needed [2].
Some of the main technologies used in DG are photovoltaic system, wind power, fuel cells,
microturbines and diesel generators. Each technology has limitation in their application and
operation that makes them more or less suitable to meet the various aim of installing DG.
11 | P a g e

12. Statement Of The Problem
DG possesses inherent advantages, conversely it’s not without disadvantages. As a result, DG
interconnection results in operating situation which does not occur in centralized power systems.
These operating situation present unique engineering challenges to DG interconnection.
This project deals with this particular operating situation that occurs at the interconnection or
Point of Common Coupling (PCC) between DG plant and the rest of the power system in the
event of a faulted condition, a situation hereafter refer to as Islanding.
One of the new technical issues created by DG interconnection is unintentional islanding.
Islanding occurs when a portion of the distribution system becomes electrically isolated from
the remainder of the power system, yet continues to be energized by DG connected to the
isolated subsystem (shown in Figure 3).
The island is an unregulated power system. Its behavior is unpredictable due to the
power mismatch between the load and generation and the lack of voltage and frequency
control. The main concerns associated with such islanded systems are: [21]
 The voltage and frequency provided to the customers in the islanded system can vary
significantly if the distributed generators do not provide regulation of voltage and
frequency and do not have protective relaying to limit voltage and frequency excursions,
since the supply utility is no longer controlling the voltage and frequency, creating the
possibility of damage to customer equipment in a situation over which the utility has no
control. Utility and DG owners could be found liable for the consequences.
 Islanding may create a hazard for utility line-workers or the public by causing a line to
remain energized that may be assumed to be disconnected from all energy sources.
 The distributed generators in the island could be damaged when the island is
reconnected to the supply system. This is because the generators are likely not in
synchronism with the system at the instant of reconnection. Such out-of-phase reclosing
12 | P a g e

13. can inject a large current to the generators. It may also result in re-tripping in the supply
system. [21]
 Islanding may interfere with the manual or automatic restoration of normal service for
the neighboring customers. [21]
It can be desirable to permit such islanded operation to increase customer reliability, and this is
often done where the DG provides backup power to the facility where it is installed. However,
considerable engineering effort, control functionality, and communications infrastructure are
necessary to make intentional islanding viable where the island includes a portion of primary
system and other loads. Even greater requirements are necessary to coordinate the operation of
more than one DG in an island. In general, if provision has not been made for islanded operation
beyond the local facility load, any unintentional islands which do occur are undesired.
Typically, according to IEEE 1547 a DG should be disconnected within 100 to 300 ms after loss
of main supply [1]. Hence there’s need to quickly detect and eliminate unintentional DG
supported islands in the event of a faulted condition. Ideally, the fault should be detected by the
DG protection system and the DG tripped before the formation of an island. To achieve this each
distributed generator must be equipped with an islanding detection device or anti islanding
devices, such as, vector shift relay and ROCOF relay [4] [5].
Before After
Figure 3: Utility Network before and after islanding has occurred.
13 | P a g e

14. SCOPE OF WORK
Overview
This project will involve examining the national utility (GPL) network (single line diagrams) to
identify potential unintentional islanding conditions, subsequently an equivalent of all the
portions of the network with potential for islanding will be produced (in the form of a single line
diagram). This equivalent single line diagram of the network containing the island/s will then be
used to develop a Matlab/Simulink model. The model will contain anti islanding relays, such as,
rate of change of frequency (ROCOF), vector surge, over/under voltage relays, over/under
current relays and over/under frequency relays based on the principal governing their operation.
The model will be simulated under a predefined or intentional islanding condition, so as to
evaluate and determine the performance of these relays for the purpose of assisting electrical
protection engineers in selecting the most appropriate protective devices and their corresponding
settings for DG systems.
14 | P a g e

15. Literature Review
Rationale for anti-islanding protection:
Anti-islanding capability is an important requirement for distributed generators. It refers to the
capability of a distributed generator to detect if it operates in an islanded system and to
disconnect itself from the system in a timely fashion. Failure to trip islanded generators can lead
to a number of problems for the generator and the connected loads. The current industry practice
is to disconnect all distributed generators immediately after the occurrence of islands.
The main philosophy of detecting an islanding situation is to monitor the DG output parameters
and system parameters, and based on system requirements whether or not an islanding situation
has occurred from change in these parameters. Islanding detection techniques can be divided into
remote and local techniques and local techniques can further be divided into passive, active and
hybrid techniques as shown in Figure 4 [5].
Islanding
Detection
Remote Local
Technique Technique
Power Line
Transfer Trip Passive Active Hybrid
Signaling
Scheme Technique Technique Technique
Scheme
Positive
Rate of Change Reactive Power Voltage and
Voltage/Current Rate of Change Vector Surge Phase/Frequency Feedback and
of Output Export Error Reactive Power
Unballance of Frequency Detection Shift Method Voltage
Power Detection Shift
Imballance
Under/Over Under/Over
Voltage Current
Figure 4: Islanding detection techniques.
15 | P a g e

16. Remote Islanding Detection Techniques
Remote islanding detection techniques are based on communication between utilities and DGs.
Although these techniques may have better reliability than local techniques, they are expensive
to implement and hence uneconomical .Some of the remote islanding detection techniques are as
follows:
a) Power line signaling scheme
These methods use the power line as a carrier of signals to transmit islanded or non-islanded
information on the power lines. The apparatus includes a signal generator at the substation that is
coupled into the network where it continually broadcasts a signal. Each DG is then equipped
with a signal detector to receive this transmitted signal. Under normal operating conditions, the
signal is received by the DG and the system remains connected. However, if an island state
occurs, the transmitted signal is cut off because of the substation breaker opening and the signal
cannot be received by the DG, hence indicating an island condition [4] [5].
This method has the advantages of its simplicity of control and its reliability. However there are
also several significant disadvantages to this method, the fist being the practical implementation.
To connect the device to a substation, a high voltage to low voltage coupling transformer is
required. A transformer of this voltage capacity can be very expensive.
Another problem for power line communication is the complexity of the network and the
affected networks. A perfectly radial network with one connecting breaker is a simple example
of island signaling; however, more complex systems with multiple utility feeders may find that
differentiation between upstream breakers difficult [5].
b) Transfer trip scheme:
The basic idea of transfer trip scheme is to monitor the status of all the circuit breakers and
reclosers that could island a distribution system. Supervisory Control and Data Acquisition
(SCADA) systems can be used for that. When a disconnection is detected at the substation, the
transfer trip system determines which areas are islanded and sends the appropriate signal to the
16 | P a g e

17. DGs, to either remain in operation, or to discontinue operation. Transfer trip has the distinct
advantage similar to Power Line Carrier Signal that it is a very simple concept. With a radial
topology that has few DG sources and a limited number of breakers, the system state can be sent
to the DG directly from each monitoring point [5] [6].
The weaknesses of the transfer trip system are better related to larger system complexity cost and
control. As a system grows in complexity, the transfer trip scheme may also become obsolete,
and need relocation or updating. The other weakness of this system is control. As the substation
gains control of the DG, the DG may lose control over power producing capability. If the transfer
trip method is implemented correctly in a simple network, there are no non-detection zones of
operation.
Local Detection Techniques
It is based on the measurement of system parameters at the DG site, like voltage, frequency, etc.
It is further divided into passive, active and hybrid detection technique.
1. Passive detection techniques
Passive methods work on measuring system parameters such as variations in voltage, frequency,
harmonic distortion, etc. These parameters vary greatly when the system is islanded.
Differentiation between an islanding and grid connected condition is based upon the thresholds
set for these parameters. Special care should be taken while setting the threshold value so as to
differentiate islanding from other disturbances in the system. Passive techniques are fast and they
don’t introduce disturbance in the system but they have a large non detectable zone (NDZ) where
they fail to detect the islanding condition [4] [5].
There are various passive islanding detection techniques and some of them are as follows:
a) Rate of change of output power
𝑑𝑝
𝑑𝑡
The rate of change of output power, , at the DG side, once it is islanded, will be much greater
than that of the rate of change of output power before the DG is islanded for the same rate of
17 | P a g e

18. load change[7]. It has been found that this method is much more effective when the distribution
system with DG has unbalanced load rather than balanced load. [5] [8]
b) Rate of change of frequency
𝑑𝑓
𝑑𝑡
The rate of change of frequency, , will be very high when the DG is islanded. The rate of
change of frequency (ROCOF) can be given by [9]
= ∗ 𝑓
𝑑𝑓 ∆𝑝
𝑑𝑡 2𝐻𝐺
ROCOF:
Where ∆𝑝 is the power mismatch at the DG side.
H is the moment of inertia for the DG/system.
G is the rated generation capacity of the DG/system.
𝑑𝑓
Large systems have large H and G where as small systems have small H and G giving larger
𝑑𝑡
value for ROCOF relay monitors the voltage waveform and will operate if ROCOF is
higher than setting for certain duration of time. The setting has to be chosen in such a way that
the relay will trigger for island condition but not for load changes. This method is highly
reliable when there is large mismatch in power but it fails to operate if DG’s capacity matches
with its local loads [5].
Figure 5: Equivalent Circuit of Synchronous Generator equipped with ROCOF Relay
operating parallel with Utility [4,6].
18 | P a g e

19. Figure 5 presents an equivalent circuit of a synchronous generator equipped with a ROCOF
relay operating in parallel with a distribution network. In this figure, a synchronous generator
(SG) feeds a load (L). The difference between the electrical powers PSG supplied by the
generator and PL consumed by the load is provided (or consumed) by the main grid.
Therefore, the system frequency remains constant. If the circuit breaker (CB) opens, due to a
fault for example, the system composed by the generator and the load becomes islanded.
In this case, there is an electrical power imbalance due to the lost grid power PSYS This power
imbalance causes transients in the islanded system and the system frequency starts to vary
dynamically. Such system behavior can be used to detect an islanding condition. However, if
the power imbalance in the islanded system is small, then the frequency will change slowly.
Thus, the rate of change of frequency can be used to accelerate the islanding detection for this
situation. [4, 5] The rate of change of frequency is calculated considering a measure window
over a few cycles, usually between 2 and 50 cycles.
This signal is processed by filters and then the resulting signal is used to detect islanding. If the
value of the rate of change of frequency is higher than a threshold value, a trip signal is
immediately sent to the generator CB. Typical ROCOF settings installed in 60-Hz systems are
between 0.10and 1.20 Hz/s. Another important characteristic available in these relays is a block
function by minimum terminal voltage. If the terminal voltage drops below an adjustable level
Vmin , the trip signal from the ROCOF relay is blocked. This is to avoid, for example, the
actuation of the ROCOF relay during generators start-up or short circuits. [5]
c) Vector Shift Detection
Vector Shift relay measures the change of phase angle of the voltage waveform to a known
reference waveform. When an island state occurs, there can be an immediate phase shift by the
DG to accommodate the change in power requirements. Once again, a threshold is set at the
maximum phase jump allowed and if the DG system exceeds that threshold, the relay is
triggered. [22]
A synchronous generator equipped with a VS relay operating in parallel with a distribution
network is depicted in Figure 6.
19 | P a g e

20. Figure 6: Equivalent circuit of Synchronous Generator equipped with Vector Shift Relay
operating parallel with Utility [4,6].
There is a voltage drop V between the terminal voltage V T and the generator internal voltage E I
due to the generator current I SG passing through the generator reactance X d . Consequently, there
is a displacement angle between the terminal voltage and the generator internal voltage, whose
phasor diagram is presented in Fig. 7(a). In Fig. 6, if the CB opens due to a fault, for example,
the system composed by the generator and the load L becomes islanded. At this instant, the
synchronous machine begins to feed a larger load (or smaller) because the current I SYS provided
(or consumed) by the power grid is abruptly interrupted. Thus, the generator begins to decelerate
(or accelerate).
Therefore, the angular difference between V T and E I is suddenly increased (or decreased) and
the terminal voltage phasor changes its direction, as shown in Fig. 7(b). Analyzing such
phenomenon in the time domain we see that the instantaneous value of the terminal voltage
jumps to another value and the phase changes as depicted in Fig. 8, where the point ‘A’ indicates
the islanding instant. Additionally, the frequency of the terminal voltage also changes. This
behavior of the terminal voltage is called vector shift. VS relays are based on such phenomena.
VS relays available in the market measure the duration time of an electrical cycle and start a new
measurement at each zero rising crossing of the terminal voltage. The current cycle duration
(measured waveform) is compared with the last one (reference cycle). In an islanding situation,
the cycle duration is either shorter or longer, depending on if there is an excess or a deficit of
active power in the islanded system, as shown in Fig. 8.
20 | P a g e

21. This variation of the cycle duration results in a proportional variation of the terminal voltage
angle, which is the input parameter of VS relays. If the variation of the terminal voltage angle
exceeds a predetermined threshold, a trip signal is immediately sent to the CB. Usually, VS
relays allow this angle threshold to be adjusted in the range from 2 to 20. The relay is also
disabled if the magnitude of the terminal voltage drops below a threshold value to avoid false
operation.
Figure 7: Internal and terminal voltage phasors (a) before opening with CB (b) after opening
with CB.
Figure 8: Voltage Vector Surge.
21 | P a g e

22. d) Voltage unbalance
Once the islanding occurs, DG has to take change of the loads in the island. If the change in
loading is large, then islanding conditions are easily detected by monitoring several parameters:
voltage magnitude, current magnitude, and frequency change. However, these methods may not
be effective if the changes are small. As the distribution networks generally include single-phase
loads, it is highly possible that the islanding will change the load balance of DG. Furthermore,
even though the change in DG loads is small, voltage unbalance will occur due to the change in
network condition [11] [12].
Under/Over Voltage
Under and over voltage are also used for passive islanding detection, and often as a
complementary device coupled with frequency monitoring. Voltage variations occur as a result
of a mismatch of reactive power. This relay operates on the principle that an excess of reactive
power mismatch will drive the voltage up and a deficit of reactive power will drive the voltage
down. Once the voltage falls out of the preset thresholds, the relay will open the breaker.
Hence, by determining the voltage change or its rate of change, it is possible to detect island
states that frequency effects alone cannot. Unfortunately, there is limited experience indicating
that the reactive power measurement relay will have higher performance than frequency
variations. As real power draw is often much greater than reactive power, a loss of mains is more
likely to significantly change the active power than the reactive power.
22 | P a g e

23. 2. Active detection techniques
With active methods, islanding can be detected even under the perfect match of generation and
load, which is not possible in case of the passive detection schemes. Active methods directly
interact with the power system operation by introducing perturbations. The idea of an active
detection method is that this small perturbation will result in a significant change in system
parameters when the DG is islanded, whereas the change will be negligible when the DG is
connected to the grid.
a) Reactive power export error detection
In this scheme, DG generates a level of reactive power flow at the point of common coupling
(PCC) between the DG site and grid or at the point where the Reed relay is connected [14] [15].
This power flow can only be maintained when the grid is connected. Islanding can be detected if
the level of reactive power flow is not maintained at the set value. For the synchronous generator
based DG, islanding can be detected by increasing the internal induced voltage of DG by a small
amount from time to time and monitoring the change in voltage and reactive power at the
terminal where DG is connected to the distribution system. A large change in the terminal
voltage, with the reactive power remaining almost unchanged, indicates islanding. [16]The major
drawbacks of this method are it is slow and it cannot be used in the system where DG has to
generate power at unity power factor.
b) Phase (or frequency) shift methods
Measurement of the relative phase shift can give a good idea of when the inverter based DG is
islanded. A small perturbation is introduced in form of phase shift. When the DG is grid
connected, the frequency will be stabilized. When the system is islanded, the perturbation will
result in significant change in frequency. The Slip-Mode Frequency Shift Algorithm (SMS) uses
positive feedback which changes phase angle of the current of the inverter with respect to the
deviation of frequency at the PCC. A SMS curve is designed in such a way that its slope is
greater than that of the phase of the load in the unstable region. [5] [17]
23 | P a g e

24. The drawback of this method is that the islanding can go undetected if the slope of the phase of
the load is higher than that of the SMS line, as there can be stable operating points within the
unstable zone [18].
3. Hybrid detection schemes
Hybrid methods employ both the active and passive detection techniques. The active technique is
implemented only when the islanding is suspected by the passive technique. Some of the hybrid
techniques are as follows:
a) Technique based on positive feedback (PF) and voltage imbalance (VU)
This islanding detection technique uses the PF (active technique) and VU (passive technique).
The main idea is to monitor the three-phase voltages continuously to determinate VU which is
𝑣 + 𝑠𝑞
given as
𝑉𝑈 =
𝑣 − 𝑠𝑞
V+Sq and V-Sq are the positive and negative sequence voltages, respectively. Voltage spikes
will be observed for load change, islanding, switching action, etc. Whenever a VU spike is
above the set value, frequency set point of the DG is changed. The system frequency will
change if the system is islanded [19].
b) Technique based on voltage and reactive power shift
In this technique voltage variation over a time is measured to get a covariance value (passive)
which is used to initiate an active islanding detection technique, adaptive reactive power shift
(ARPS) algorithm [20].
instead of current phase shift. The d-axis current shift, 𝑖 𝑑 or reactive power shift is given as
𝑘
The ARPS uses the same mechanism as ALPS, except it uses the d-axis current shift
24 | P a g e

25. 𝑇 𝑎𝑣′ − 𝑇 𝑣
(𝑘)
𝑖 = 𝑘𝑑�
𝑘
�
𝑇𝑣
𝑑 (𝑘)
Where;
Tav' is the average of the previous four voltage periods.
Uav is the mean of Tav'
Tv is the voltage periods
UV is the mean of TV
kd is chosen such that the d-axis current variation is less than 1 percent of q-axis current in
inverter's normal operation. The additional d-axis current, after the suspicion of island, would
accelerates the phase shift action, which leads to a fast frequency shift when the DG is islanded
[5].
25 | P a g e

26. METHODS/DESIGN APPROACH
Detection of Islanded Power Systems
An islanding situation should be detected soon after the island is formed. The basic
requirements for a successful detection are:
 The scheme should work for any possible formations of islands. Note that there could be
multiple switchers, reclosers and fuses between a distributed generator and the supply
substation. Opening of any one of the devices will form an island. Since each island
formation can have different mixture of loads and distributed generators, the behavior of
each island can be quite different. A reliable anti-islanding scheme must work for all
possible islanding scenarios.
 The scheme should detect islanding conditions within the required time frame. The main
constraint here is to prevent out-of-phase reclosing of the distributed generators. A
recloser is typically programmed to reenergize its downstream system after about 0.5 to
1 second delay. Ideally, the anti-islanding scheme must trip its DG before the reclosing
takes place.
26 | P a g e

27. Network Studied
A detail examination of GPL’s DIS revealed that there were at least three cases that possess potential for
islanding. Of these three cases, the Versailles/Lenora area was chosen to carry out the study, since the
only DGs present in the DIS was found to be located in this area.
From Garden Of Eden
Figure 9: One line diagram for the Versailles/Lenora portion of GPL’s DIS.
The one line diagram in Figure 9 shows the Versailles/Lenora portion of the DIS, and more so,
the area of interest. However from visits made to Versailles it was found out that changes were
made to the system that was not documented or updated in the one line diagram, changes such
as, the ‘A1’ and ‘A3’ generator sets were no longer operational and there were three generators
sets present at Lenora instead of two show in the diagram. Hence taking the network
configuration as shown in the Figure 9 and the changes that were made to the system, a modified
equivalent diagram was produced (shown in Figure 10).
27 | P a g e

28. CB1
CB2
Figure 10: Modified Equivalent One Line Network Diagram.
Since the four mobile Caterpillar sets (A2, A4,A7 and A8) at Versailles and the three (A1,A2
and A3) at Lenora all are the same model, that is, all having the same parameters, for connivance
they were combined and model as a single generator at each location. The equivalent one line
diagram is made up of the following:
Generator G1 (A6) Generator G2 (A2, A4, A7 & A8)
Model#: GM AB20-24 4 Mobile Caterpillar Generator set
3250 KVA 60Hz 4160V Model #: 3516
2000 KVA 60Hz 480V
Generator G3 (A1 & A2) Model #: 3516
3 Mobile Caterpillar Generator set 2000 KVA 60Hz 480V
28 | P a g e

29. Transformer T1
3750KVA 4160/13800V Δ/Υ
Transformer T2 (4) & T3 (3) Garden of Eden (GOE) Interconnection
2500KVA 480/13800V Δ/Υ 3 phase current Source
Feeders
West Bank approx 5 MW
West Coast approx 9 MW
In creating the model shown in figure 11, it was found that Versailles and Lenora together have a
generating capacity of 17.25 MVA (11.25 MVA at Versailles and 6.0 MVA at Lenora) plus
power imported from GOE which is approximately 5.0 MVA, hence this gives a total of 22.25
MVA. However the total load demand of the two feeders (west coast and west bank) connected
to Versailles and Lenora was found to be approximately 14 MVA, thus having a surplus of 8.25
MVA in generation. Therefore an assumption was made that all the generators either at
Versailles or at Lenora was not in operation at the same time, hence this was taken into
consideration when creating the model, that is, instead of combining all four of the mobile
Caterpillar sets at Versailles only two was combined and modeled to produce 4 MVA, however,
all three at Lenora was model as being in operation.
29 | P a g e

30. Simulation model
In order to investigate the performance of the different techniques used in island detection a
simulation model was implemented. The model is based upon a specific portion of GPL’s
Demerara Interconnected System (DIS) (shown in figure 10), and was created so that the model
reflects the real system as much as possible. The behavior of the simulated system must be
similar to what happens in a real situation.
Figure 11: Matlab/Simulink model of Versailles and Lenora.
Figure 11 shows the Matlab/Simulink model for the area of interest (Versailles and Lenora) and
is based on the equivalent one line diagram depicted in figure 10.
The model contain three synchronous generator (G1, G2 & G3), three transformers (T1, T2 &
T3), two circuit breakers, a three phase source representing Garden Of Eden (GOE), one feeder
for the west bank and one for the west coast each consuming 5 MW and 9 MW respectively and
various monitoring and measuring blocks.
30 | P a g e

31. The synchronous generated G1 is rated at 3250 KVA 60Hz 4160 V, G2 is a combination of four
mobile Caterpillar sets each rated at 2000 KVA 60Hz 480 V and G3 is a combination of three
mobile Caterpillar sets all with same ratings used for G2.
The transformer T1 is rated at 3750KVA 4160/13800V Δ/Υ, T2 is a combination of four
individual transformer each with ratings of 2500KVA 480/13800V Δ/Υ and T3 is a combination
three individual transformer all with same rating as ones in T2.
The three phase source representing GOE contribution to Versailles is based on the maximum
short circuit current level during a line-to-ground fault between Versailles and GOE
interconnection, multiplied by the line-to-line voltage (VA).
Note: For all combination of generators and transformers, the rated power is summed and all
impedances are parallel.
Conditions for Islanding
There are basically two conditions for islanding in the network studied. These conditions are:
1. When ‘CB1’ (circuit breaker 1) depicted in figure 10 is in the open position. That is, it
disconnects the entire Versailles and Lenora from the rest of the DIS forming what we
may refer to hereafter as a major island. The term major island is used because Versailles
location is not considered to be distributed generation but since its at the end of the DIS
where there’s only one interconnection from Versailles to the rest of the grid, any
disruption in this connection can leave it isolated from the rest of the DIS and hence
islanded. Also Lenora DG’s would also be considered to be a part of the Versailles Island
as shown in figure 12.
2. When ‘CB2’ (circuit breaker 2) depicted in figure 10 is in the open position. Since it
disconnect the DG’s at Lenora from the rest of the grid thus forming a minor island. A
minor island since the Lenora location meets the criteria of being distribution generation
and the resulting island will only be made up of the generators at Lenora.
31 | P a g e

32. G1
T1
CB1
B1
G2 ~
CB2
3 Phase Source
GOE
T2
West Bank
Major Island
G3
T3
Minor Island
B2
West Coast
Figure 12: Distinction between Major and Minor Island and Conditions for Islanding in the
Network Studied.
32 | P a g e

33. Model Description
Synchronous Machine (Alternator, Diesel Engine Speed & Voltage Control)
Figure 13: Simulink model of a Synchronous Machine
The Synchronous Machine block operates in generator or motor modes. The operating mode is
dictated by the sign of the mechanical power (positive for generator mode, negative for motor
mode). The model takes into account the dynamics of the stator, field, and damper windings. The
equivalent circuit of the model is represented in the rotor reference frame (q&d frame). All rotor
parameters and electrical quantities are viewed from the stator.
The SM voltage and speed outputs are used as feedback inputs to the diesel engine speed &
voltage control block which contains governor block as well as an excitation block.
Three Phase Transformer Block (Two Winding)
R1 L1 L2 R2
Rm Lm
Figure 14: Simulink model of a three phase transformer and its equivalent circuit.
This block implements a three-phase transformer using three single-phase transformers. The
Linear Transformer block model shown consists of two coupled windings wound on the same
core. The model takes into account the winding resistances (R1 and R2) and the leakage
33 | P a g e

34. inductances (L1 and L2), as well as the magnetizing characteristics of the core, which is modeled
by a linear branch (Rm Lm).
The two windings of the transformer can be connected as follows:
 Y
 Y with accessible neutral
 Grounded Y
 Delta (D1), delta lagging Y by 30 degrees
 Delta (D11), delta leading Y by 30 degrees
Three-Phase Source
Figure 15: Simulink model of a three phase source.
Implement three-phase source with internal R-L impedance.
The Three-Phase Source block implements a balanced three-phase voltage source with internal
R-L impedance. The three voltage sources are connected in Y with a neutral connection that can
be internally grounded or made accessible. You can specify the source internal resistance and
inductance either directly by entering R and L values or indirectly by specifying the source
inductive short-circuit level and X/R ratio.
Note: For the model that was created, a three phase source was used to model Garden of Eden
interconnection to Versailles and more so Versailles interconnection to the entire grid. For this
source the short circuit level (VA) and X/R ratio was specified.
34 | P a g e

35. Three-Phase Breaker
Figure 16: Simulink model of a three phase breaker.
The Three-Phase Breaker block implements a three-phase circuit breaker where the opening and
closing times can be controlled either from an external Simulink signal (external control mode),
or from an internal control timer (internal control mode).
The Three-Phase Breaker block uses three Breaker blocks connected between the inputs and the
outputs of the block. You can use this block in series with the three-phase element you want to
switch.
If the Three-Phase Breaker block is set in external control mode, a control input appears in the
block icon. The control signal connected to this input must be either 0 or 1, 0 to open the
breakers, 1 to close them. If the Three-Phase Breaker block is set in internal control mode, the
switching times are specified in the dialog box of the block. The three individual breakers are
controlled with the same signal.
Three-Phase Parallel RLC Load
Figure 17: Simulink model of a three phase parallel RLC load.
The Three-Phase Parallel RLC Load block implements a three-phase balanced load as a parallel
combination of RLC elements. At the specified frequency, the load exhibits constant impedance.
The active and reactive powers absorbed by the load are proportional to the square of the applied
voltage.
35 | P a g e

36. RMS Block-
Figure 18: Simulink model of a root mean square (rms) calculation block.
This block measures the true root mean square value, including fundamental, harmonic, and DC
components, of an instantaneous current or voltage. The RMS value of the input signal is
calculated over a running average window of one cycle of the specified fundamental frequency,
where f(t) is the input signal and T is 1/(fundamental frequency). Since this block uses a running
average window, one cycle of simulation has to be completed before the output gives the correct
value.
The discrete version of this block allows you to specify the initial magnitude of the input signal.
For the first cycle of simulation the output is held to the RMS value of the specified initial input.
Three-Phase V-I Measurement
Figure 19: Simulink model of a three phase voltage-current measurement block.
The Three-Phase V-I Measurement block is used to measure instantaneous three-phase voltages
and currents in a circuit. When connected in series with three-phase elements, it returns the three
phase-to-ground or phase-to-phase peak voltages and currents.
The block can output the voltages and currents in per unit (pu) values or in volts and amperes.
36 | P a g e

37. If you choose to measure phase-to-ground voltages in per unit, the block converts the measured
voltages based on peak value of nominal phase-to-ground voltage:
where
If you choose to measure phase-to-phase voltages in per unit, the block converts the measured
voltages based on peak value of nominal phase-to-phase voltage:
where
If you choose to measure currents in per unit, the block converts the measured currents based on
the peak value of the nominal current:
where
V nom and P base are specified in the Three-Phase V-I Measurement block dialog box.
37 | P a g e

38. Display Block
Figure 20: Simulink model of display block.
The Display block shows value of an inputted signal. It accepts real or complex signals of the
following data types:
• Floating point
• Built-in integer
• Fixed point
• Boolean
• Enumerated
Scope Block
Figure 21: Simulink model of an oscilloscope (scope).
The Scope block displays signal inputs with respect to simulation time and displays signal
generated during the simulation.
38 | P a g e

39. Protection Block
Figure 22: Simulink model for the protection sub-system.
The protection block is a sub-system that contains all the protection relays (shown in figure 22).
These include the under/over current relay, under/over voltage relay, under/over frequency relay,
rate of change of frequency (ROCOF) relay and the vector shift relay.
Figure 23: Relays found inside the protection sub-system block
39 | P a g e

40. Each relay is equipped with two display, one which indicate the status of the relay (‘1’ indicates
a trip status) and the other displays and log the time at which the relay was activated or trip. Each
relay is only activated once during the entire simulation, that is, at the first instance to which it
senses an abnormal condition or a condition to which it was designed to sense/activate.
The following is a detail description of all the relays contained in the protection block and their
corresponding setting.
Under/Over Current, Under/Over Voltage and Under/Over Frequency Relay
Figure 24: (a) Simulink model of Under/Over Current Relay model.
40 | P a g e

41. Figure 24: (b) Simulink model of Under/Over Voltage Relay model.
Figure 24: (c) Simulink model of Under/Over Frequency Relay model.
41 | P a g e

42. Figure 24 (a) depicts the Matlab/Simulink model of the under/over current relay. The operation
of the model is based on the actual relay, where the line current (all three phases) of the system is
monitored and compared to some preset thresholds (a maximum value for over current and
minimum value for under current) and if the line current goes over or under these thresholds for
a predetermined period then a trip signal is initiated.
In the model, the line current (Iabc) is continually monitor and compared to the set thresholds,
this comparison is done by using a ‘Relational Operator’, that is, instances where the line current
is greater than (>) or less than (<) the maximum or minimum current value respectively, it
outputs a signal to the On/Off delay. If this signal (output from the relational operator) remains at
the input of the on/off delay for longer than the preset time a ‘trip’ signal is initiated and the time
for which the trip signal was initiated is logged and display. It’s important to note that since an
‘OR’ gate was used at the output of the relay, for an abnormal condition in any of the three
phases, a trip signal is initiated.
The under/over voltage relay and the under/over frequency relay shown in figure 24 (b) and
figure 24 (c) respectively operates on the same principle as the under/over current relay. In the
case of the under/over voltage relay the only difference is that the parameter in which the relay
monitors, that is, the under/over voltage relay monitors the three phase voltage, while the
under/over current relay monitors three phase current and similarly the under/over frequency
relay monitors the system frequency. However the setting of these relay will be different from
each other.
42 | P a g e

43. The Rate of Change of Frequency Relay
Figure 24 (d): Simulink model of The Rate of Change of Frequency Relay (ROCOF).
Figure 24 (d) shows the Simulink model of the rate of change of frequency relay. Unlike the rest
of relay model describe thus far, the ROCOF relay accepts or monitor two inputs (frequency and
terminal voltage Vt), therefore before the relay is activated two conditions must be satisfied. The
frequency with time � �, the absolute value of the rate of change of frequency � � it is then
𝑑𝑓 𝑑𝑓
frequency is fed into a ‘Discrete Derivative’ block which calculates the rate of change of
𝑑𝑡 𝑑𝑡
compared (using a Relational Operator) to the ROCOF threshold and if it exceeds this threshold,
the output of the relational operator goes ‘true’ and a ‘1’ is sent into the first input of the ‘AND’
gate. However before the ‘AND’ gate can output a signal to initiate a trip, another condition must
be met, that is, the terminal voltage Vt (pu) of the generator is compared to a set threshold and if
it exceeds this threshold the second input of the ‘AND’ gate goes ‘true’ (that is both condition is
satisfied), hence the output of the ‘AND’ gate also goes ‘true’ which immediately starts the delay
43 | P a g e

44. count down. If the ‘AND’ output remains ‘true’ for longer period than a predetermine time (set
by the On Delay) then and only then a trip signal is initiated and the time of the trip is logged.
Vector Shift Relay
Figure 24 (e): Simulink model of a Vector Shift Relay.
Figure 24 (e) depicts the Simulink model of a vector shift relay. Similar to the ROCOF relay, the
vector shift relay also accepts or monitors two inputs (three phase voltage Vabc and terminal
voltage Vt), and therefore two conditions must also be satisfied before the relay can activate.
The relay monitors the three phase waveform and counts every complete cycle by detecting the
rising edge of the wave, and at the same time the duration of each cycle or the period is
measured. Since the model operates at frequency (f) of 60 Hz, therefore the period (T) will be
44 | P a g e

45. equal to (1/f) 0.01667 seconds. Hence the model computes the duration of each period by
dividing the cycle time by the number of completed period/s and then compares this value to
0.01667 seconds and any time value grater or less than the set threshold the first condition is
reached. But before a trip is initiated the second condition must be met, that is, if the terminal
voltage (Vt) exceeds the set threshold and both conditions are met then and only then a trip
signal is sent and the time is logged.
45 | P a g e

46. SIMULATION RESULTS
Normal Conditions
The complete model was simulated at normal condition for 5 seconds and the results obtained
are shown below.
Note: Since the purpose of the simulation is to compare performance of the relays with respect to
time, a small sampling time was chosen for the simulation, more specifically 50 micro seconds.
Therefore 5 seconds will be more than adequate for the entire simulation run time.
Figure 25 (a), (b) and (c) shows the results obtained from the three synchronous generators used
in the model. In each figure, the mechanical power input (Pmec), excitation voltage (Vf),
terminal voltage (Vt) and speed all in per-unit is displayed. From looking at all three of the
figures obtained for the generators (SM1, SM2 and SM3), it can be clearly seen that they
system/model initially takes approximately 1 second to reach a steady state condition. Using the
graphs, the steady state values of Pmec, Vf, Vt and speed can be approximated to the following:
Steady State Approximated Values (pu)
Synchronous Machines
Pmec Vf Vt Speed
SM1 0.255 1.500 1.000 1.000
SM2 0.315 0.910 1.000 1.000
SM3 0.289 1.360 1.000 1.000
Table 1: Steady state parameters for the Synchronous Machines used in the model, under normal
operating conditions.
46 | P a g e

47. Figure 25 (a): Simulation result of Synchronous Machine One (SM1).
Figure 25 (b): Simulation result of Synchronous Machine Two (SM2).
47 | P a g e

48. Figure 25 (c): Simulation result of Synchronous Machine Three (SM3).
Figure 26 (a) and (b) shows the 3 phase voltages and currents at Bus 1 and 2 found in the
system/model. The first figure (fig 26 a) shows the voltages and current for a three (3) seconds
period after the simulation was started. It can be seen that the voltages Vabc at Bus 1 (i.e.
Vabc_B1) and Vabc at Bus 2 (i.e. Vabc_B2) are relatively constant throughout the simulation
while the currents Iabc at Bus 1 (i.e. Iabc_B1) and Iabc at Bus 2 (i.e. Iabc_B2) takes
approximately one (1) second after the simulation has started to become constant. It can also be
seen that the voltages at both Bus is approximately the same while the currents vary in value
from each other.
Figure 26 (b) shows and expanded portion of figure 26 (a).
48 | P a g e

49. Figure 26 (a): Simulation result for 3 phase voltages and currents at bus 1 and bus 2.
Figure 26 (b): Expanded view of the 3 phase voltages and currents at bus 1 and bus 2.
49 | P a g e

50. Figure 27 (a): Simulation result for the 3 phase rms voltages at bus 1&2.
Figure 27 (b): Simulation result for the 3 phase rms currents at bus 1&2.
50 | P a g e

51. Figure 27 a, and b shows the 3 phase rms voltages and currents at bus 1 and 2 respectively, here
again it can be clearly seen that the voltages in each phase are constant (approximately 13790
volts) after about 1 second into the simulation. The currents also follow the same pattern but vary
in value at each bus, that is, the average rms value for the currents in all three phase is 61.60
Amps at bus 1 and 84.72 Amps at bus 2.
Figure 28: Simulation result of the rate of change of frequency and frequency for bus 1 & bus 2.
The final figure (fig 28) shows the frequency at bus 1 and 2 and their respective rate of change of
frequency. As expected, both the frequency and the rate of change of frequency reach a steady
state or become constant after the 1 second mark. It can also be observed that for a small change
in frequency (60 to 60.4) results in a relatively large ‘rate of change of frequency’ or large df/dt
(0 to 14).
51 | P a g e

52. Islanded Condition
Scenario 1: Formation of a Major Island (Loss of Grid)
As mention earlier, a major island is formed when the interconnection between Garden of Eden
(GOE) and Versailles is lost, hence completely isolating Versailles and Lenora (together) from
the rest of the DIS or grid, a scenario which arises when CB1 (shown in figure 12) is in the open
position.
To achieve this scenario and for purpose of this project, CB1 was pre configured to open on all
three phase, 3 seconds after the simulation was started hence forming a major island to illustrate
the effects that an unintentional island has on a power distribution network.
Hence the following results were obtained from this simulated scenario.
Figure 29 (a): Simulation result for Synchronous Machine One (SM1) for a major islanded
condition.
52 | P a g e

53. Figure 29 (b): Simulation result for Synchronous Machine One (SM2) for major islanded
condition.
Figure 29 (c): Simulation result for Synchronous Machine One (SM3) for major islanded
condition.
53 | P a g e

54. Figure 29 a, b, and c shows the simulation results three synchronous generators, SM1, SM2 and
SM3 under a major islanded condition. All three generators basically responded to the islanded
or loss of grid condition in similar manner, that is, after the island was formed (3 seconds into the
simulation), there were an immediate increase in mechanical power (Pmec) supplied to the
generator since due to the loss in grid the three generator had to supplied the required power
demand on their own, hence there were an increase in load to each generator. Also to counteract
this increase there were also an increase in excitation voltage (Vf) to the alternator, we can also
see that the terminal voltage (Vt) and speed of the generator was also affected by an increase
load at each generator.
Figure 30 (a): Simulation result for the 3 phase rms currents at bus 1&2 for major islanded
condition.
54 | P a g e

55. Figure 30 (b): Simulation result for the 3 phase rms voltages at bus 1&2 for a major islanded
condition.
Figure 31: Simulation result for the 3 phase voltages and current at bus 1&2 for a major islanded
condition.
55 | P a g e

56. Figure 31 shows the three phase voltages and currents at bus 1 and 2 and the effect that an
islanding condition have on these values. Here we see a change in voltages or voltage wave form
(highlighted in fig 31) at the instant when the island was formed (3 seconds into the simulation).
We can also see a significant change in the currents or current waveform at both buses.
Figure 32: Simulation result of the rate of change of frequency and frequency at bus 1 and bus 2
during a major islanded condition.
The above figure shows the effect that an islanded condition has on the frequency of a power
system or distribution network. From the figure we can see that the formation of the island had
the same effect on the frequency at both the bus. At the instant where the island was formed, we
can observer that there was a sharp decline in frequency, that is, the frequency drop from 60 Hz
to about 58 Hz in very short time (approximately 0.25 second). The figure also shows the
corresponding rate of change of frequency (df/df) for this change in frequency, where a change
of 60 Hz to 58 Hz corresponds to a df/dt of -12 Hz/s (where the minus sign indicates a drop in
frequency).
56 | P a g e

57. Figure 33 (a): Results obtained from the protection block at bus 1 for a major islanded condition.
57 | P a g e

58. Figure 33 (b): Results obtained from the protection block at bus 2 for a major islanded condition.
58 | P a g e

59. BUS 1 BUS 2
Relays Relays
Trip Status Trip Time (s) Detection Time (s) Trip Status Trip Time (s) Detection Time (s)
ROCOF 1 3.068 0.068 ROCOF 1 3.057 0.057
Vector Shift 1 3.073 0.073 Over Current 1 3.107 0.107
Over Current 1 3.120 0.120 Under Frequency 1 3.135 0.135
Under Frequency 1 3.135 0.135 Over Voltage 1 3.144 0.144
Over Voltage 1 3.144 0.144 Under Voltage 1 3.477 0.477
Under Voltage 1 3.477 0.477 Vector Shift 1 4.175 1.175
___ ___ ___ ___
Under Current 0 Under Current 0
___ ___ ___ ___
Over Frequency 0 Over Frequency 0
Table 2: Combine results for Relay Protection Blocks 1 and 2 for a major islanded condition.
The table above shows the results obtained for the relay protection blocks 1 and 2 monitoring
buses 1 and 2 respectively. The table shows the ‘trip statuses’, ‘trip time’ and the ‘island
detection time’ for each relay. The trip status is represented by either a ‘1’ or a ‘0’, where ‘1’
indicates a trip or relay activation and a ‘0’ represent no detection. Since the island condition
occurred exactly three seconds into the simulation, the trip time shows the time elapse after the
specific relays were activated and finally the detection time shows the time take for the relay to
respond to the island condition in ascending order.
Comparing the relays performance by detection time, where the shortest time taken to detect the
island condition the greater the performance we see that at both busses or at both protection
block the ROCOF relay out performs the others. It can also be seen that at bus 1 the relay with
the longest detection time was the under voltages relay, similarly the relay with the longest
detection time at bus 2 was the vector shift relay. And finally the under current and the over
frequency failed entirely to detect the island condition.
59 | P a g e

60. Relay Peotection at Bus 1
0.500 Relay
Performance
ROCOF
0.450
Decrease
0.400
0.350 Vector Shift
Detection Time (s)
Over Current
0.300
Under Frequency
0.250
Over Voltage
0.200
Under Voltage
0.150
0.100
0.050
Relays
0.000
Figure 34 (a): Graph showing comparison of the relays detection time at bus 1.
Relay Protection at Bus 2
1.200 Relay
Performance
ROCOF
1.000 Decrease
Over Current
Under Frequency
Detection Time (s)
0.800
0.600 Over Voltage
Under Voltage
Vector Shift
0.400
0.200
Relays
0.000
Figure 34 (b): Graph showing comparison of the relays detection time at bus 2.
60 | P a g e

61. Figure 34 (a) and (b) shows a graphical comparison of the different relays detection time, where
the performance of each relay decreases with an increase in detection time. It can be observe that
the detection time or the behavior of each relay differs depending on the location placed or the
point at which it is monitoring (i.e. bus 1 or bus 2). For example the vector shift relay was the
second relay at bus 1 to detect the island but at bus 2 it was the last, that is, at bus 1 it took 0.073
seconds to detect the island but at bus 2 it took 1.175 seconds which is approximately 16 times
longer. It can also be observed that over current, under frequency and the over voltage relays trip
in the same order at both buses but with different detection times.
61 | P a g e

62. Scenario 2: Formation of a Minor Island.
As stated earlier, a minor island is formed when the DGs’ at Lenora is disconnected or isolated
from Versailles and the rest of the DIS. Since there is only a single connection between Lenora
and Versailles, any disruption in this connection results in the formation of an island.
For the purpose of this project this scenario will be achieve by intentionally configuring CB2
(shown in figure 12) to open on all three phases 3 seconds into the simulation thus forming a
minor island and observing the effects of the island condition on the portion of the network.
Hence the following results were obtained from this simulated scenario.
Figure 35 (a): Simulation result for Synchronous Machine One (SM1) for minor islanded
condition.
62 | P a g e

63. Figure 35 (b): Simulation result for Synchronous Machine One (SM2) for minor islanded
condition.
Figure 35 (c): Simulation result for Synchronous Machine One (SM3) for minor islanded
condition.
63 | P a g e

64. Figures 35 (a), (b) and (c) shows the results obtained for the three synchronous
machines/generators (SM1, SM2, and SM3). It can be seen that SM1 and SM2 located at
Versailles was not affected much by the minor island formed at Lenora, that is, the operating
parameters (Pmec, Vf, Vt and the speed) were all maintained at an appreciable level. However
the DG at Lenora was severely affected since it was the source of the island and it was left to
supply a load that was far over its capacity. From figure 35 (c) we can see that due to an increase
in load there were an increase in mechanical power (Pmec) and excitation voltage (Vf) required
and since the DG could not have satisfied this increased load demand, the terminal voltage (Vt)
and speed decreased.
Figure 36 (a): Simulation result for the 3 phase rms currents at bus 1&2 for minor islanded
condition.
64 | P a g e

65. Figure 36 (b): Simulation result for the 3 phase rms voltages at bus 1&2 for a minor islanded
condition.
For the 3 phase rms currents and voltages at bus 1and 2 illustrated in figures 36 (a) and (b)
respectively, we see again that there were no major disturbance in voltages and current at bus 1
located at Versailles, however there were severe disturbances in the currents and voltages at bus
1 located at Lenora due to the islanded condition of the DG.
In figure 37 we can observe that there was some amount of disturbance in the frequency at bus 1,
however the extent of the disturbance would be determine by the protection block and whether it
cause a trip in any of the frequency monitoring relays. Conversely we can notice that the
frequency at bus 2 was significantly affected by the formation of this islanded condition.
65 | P a g e

66. Figure 37: Simulation result of the rate of change of frequency and frequency at bus 1 and bus 2
during a major islanded condition.
Figure 38 (a): Results obtained from the protection block at bus 1 for a minor islanded condition.
66 | P a g e

67. Figure 38 (b): Results obtained from the protection block at bus 2 for a minor islanded condition.
Figures 38 (a) and (b) shows the results obtained from protection blocks 1 and 2 respectively, at
buses 1 and 2. From figure 39 (a) it can be observed that there were no trips in any of the relays
in protection block 1 (which monitors bus 1) resulting from the islanded condition, this was
expected since the results obtained (for the minor island) showed no disturbances in the voltages,
currents or frequency. However figure 38 (b) showed multiple trips in various relays, which is
shown in details in the table below.
67 | P a g e

68. Protection Block 2/BUS 2
Relays
Trip Status Trip Time (s) Detection Time (s)
Vector Shift 1 3.035 0.035
ROCOF 1 3.036 0.036
Over Current 1 3.102 0.102
Under Frequency 1 3.115 0.115
Over Voltage 1 3.136 0.136
Under Voltage 1 4.743 1.743
Under Current 0 ___ ___
Over Frequency 0 ___ ___
Table 3: Results obtained for Relay Protection Block 2 for a minor islanded condition.
The table above shows the results obtained for the relay protections block 2 monitoring bus 2.
Here again the table shows the ‘trip statuses’, ‘trip time’ and the ‘island detection time’ for each
relay.
Comparing the relays performance by detection time, where the relay performance decreases
with the increase in detection time, we see that the vector shift relay has the shortest detection
time (0.035 seconds), that is, it was first to detect the islanded condition and the ROCOF relay
comes in second at 0.036 seconds and the last to detect the island was the under voltage relay
taking 1.743 seconds.
And here again the under current and the over frequency failed entirely to detect the island
condition.
68 | P a g e

69. Relay Protection at Bus 2
1.800
1.600
1.400
Detection Time (s)
1.200 Vector Shift
1.000 ROCOF
0.800 Over Current
0.600 Under Frequency
0.400 Over Voltage
0.200 Under Voltage
0.000
Relays
Figure 39: Graph showing comparison of the relays detection time at bus 2.
Figure 39 shows a graphical representation of the different relays detection time. Here is can be
seen that the ROCOF and the vector shift relays had the fastest detection time with a difference of
0.001 seconds. It can also be observed that over current, under frequency and the over voltage
relays trips were nearer to each other and in the same order as noticed in the first scenario or in the
major islanded condition.
69 | P a g e

70. CONCLUSION
Distributed generator interconnections near consumers have created new challenges for
protection engineers. The typical protection configurations such as unplanned islanding and
reclosing of distributed generator systems need to be address. Section 4.4.1 of the IEEE 1547
standard states: “For an unintentional island in which the DG energizes a portion of the area
electrical power system through the point of common coupling, the DG interconnection system
shall detect the island and cease to energize the Area electrical power system within one second
of the formation of an island” [1].
This thesis describes and compares different local islanding detection techniques. Fast and
accurate detection of islanding is one of the major challenges in today’s electrical power
distribution system with many distribution systems already having significant introduction of
DGs. Islanding detection is also important as islanding operation of distributed system is seen a
viable option in the future to improve the reliability and quality of the power supplied.
From the results obtained from the various simulations, it is apparent that anti-islanding relays
such as the Rate of Change of Frequency (ROCOF) and the Vector Shift relay has significant
performance with respect to detection time over traditional relays, such as, Under/Over Current,
Under/Over Voltage and Under/Over Frequency relays, where the ROCOF and Vector Shift
relays had a detection time of at least three (3) times faster that these traditional protection
relays. Also some relays (over frequency and the under current) relays failed entirely to detect
the islanded condition in both scenarios.
Consequently from the research carried out and the results/evidence provided by the
Matlab/Simulink simulations and also in keeping with international standards (more so IEEE
1547), it is of the views of the researcher that the implementation of these anti-islanding relays
(ROCOF and Vector Shift) on electrical power distribution system and more so, those containing
Distribution Generators is imperative for maintaining good quality of power and also for safe and
effective operation.
70 | P a g e

71. RECOMMENATION
The results obtained showed evidence that the Rate of Change of Frequency and the Vector Shift
relays were better at detecting the formation of island than the traditional relays that is currently
used by the national utility (GPL). It was also seen from the simulation the effects that
unintentional islanding can have on a power distribution network. Therefore to minimize these
effects and also in keeping with international standards for interconnected systems, it is therefore
recommended that these anti-islanding relays (ROCOF and Vector Shift) are implemented within
the DIS at points where may possess potentials for the formation of island.
71 | P a g e

72. BIBLIOGRAPHY
[1] “Application guide for distributed generation interconnection”. The NRECA guide to IEEE
1547, March 2006. Resource Dynamics Corporation
[2] Ann Chambers, Barry Schnoor, Stephanie Hamilton. “Distributed generation: A non technical
guide”. PennWell Corporation, 2001
[3] Gregory W. Massey. “Essentials of distributed generation systems”. Jones & Bartlett
Learning Canada. Kimberly Brophy, 2010.
[4] R. A. Walling and N. W. Miller. “Distributed generation islanding – Implications on power
system dynamic performance”.
[5] Truptimayee Pujhari. “Islanding detection in distributed generation”. M.S. thesis,
National Institute of Technology Rourkela May, 2009
[6] Ward Bower and Michael Ropp. Evaluation of islanding detection methods for
photovoltaic utility-interactive power systems. Report IEA PVPS Task 5 IEA PVPS T5-09:
2002, Sandia National Laboratories Photovoltaic Systems Research and Development, March
2002.
[7] M. A. Refern, O. Usta, and G. Fielding, “Protection against loss of utility grid supply for a
dispersed storage and generation unit,” IEEE Transaction on Power Delivery, vol. 8, no. 3, 948-
954, July 1993
[8] M. A. Redfern, J. I. Barren, and O. Usta, “A new microprocessor based islanding
protection algorithm for dispersed storage and generation, units,” IEEE Trans. Power
Delivery, vol. 10, no. 3, 1249-1254, July 1995.
[9] J. Warin, and W. H. Allen, “Loss of mains protection,” in Proc. 1990 ERA
Conference on Circuit Protection for industrial and Commercial Installation, London, UK,
4.3.1-12.
[10] F. Pai, and S. Huang, “A detection algorithm for islanding-prevention of dispersed
consumer-owned storage and generating units,” IEEE Trans. Energy Conversion, vol. 16, no. 4,
346-351, 2001.
72 | P a g e

73. [11] S. I. Jang, and K. H. Kim, “A new islanding detection algorithm for distributed
generations interconnected with utility networks,” in Proc.IEEE International Conference on
Developments in Power System Protection, vol.2, 571-574, April 2004.
[12] S. I. Jang, and K. H. Kim, “An islanding detection m e t h o d for distributed generations
using voltage unbalance and total harmonic distortion of current,” IEEE Tran. Power
Delivery, vol. 19, no. 2, 745-752, April 2004.
[13] S. Jang, and K. Kim, “Development of a logical rule-based islanding detection
method for distributed resources,” in Proc. IEEE Power Engineering Society Winter
Meeting, vol. 2, 800-806, 2002.
[14] J. Warin, and W. H. Allen, “Loss of mains protection,” in Proc. 1990 ERA Conference
on Circuit Protection for industrial and Commercial Installation, London, UK, 4.3.1-12.
[15] P. D. Hopewell, N. Jenkins, and A. D. Cross, “Loss of mains detection for small
generators,” IEE Proc. Electric Power Applications, vol. 143, no. 3, 225-230, May 1996.
[16] J. E. Kim, and J. S. Hwang, “Islanding detection method of distributed generation units
connected to power distribution system,” in Proc. 2000 IEEE Power System Technology
Conference, 643-647.
[17] G. A. Smith, P. A. Onions, and D. G. Infield, “Predicting islanding operation of grid
connected PV inverters,” IEE Proc. Electric Power Applications, vol. 147, 1-6, Jan. 2000.
[18] M. E. Ropp, M. Begovic, A. Rohatgi, G. Kern, and R. Bonn, “Determining the
relative effectiveness of islanding detection methods using phase criteria and non- detection
zones,” IEEE Transaction on Energy Conversion, vol. 15, no. 3, 290-296, Sept. 2000.
[19] V. Menon, and M. H. Nehrir, “A hybrid islanding detection technique using voltage
unbalance and frequency set point,” IEEE Tran. Power Systems, vol. 22, no. 1, 442- 448, Feb.
2007.
[20] J. Yin, L. Chang, and C. Diduch, “A new hybrid anti-islanding algorithm in grid
connected three-phase inverter system,” 2006 IEEE Power Electronics Specialists
Conference, 1-7.
73 | P a g e

74. [21] Wilsun Xu, “An Assessment of Distributed Generation Islanding Detection Methods and
Issues for Canada,” , University of Alberta Konrad Mauch, Mauch Technical Services, Sylvain
Martel, CANMET Energy Technology Centre – Varennes July 2004.
[22] Michael C. Wrinch, “Negative Sequence Impedance Measurement for Distributed Generator
Islanding Detection,” The University of British Columbia, 1995
74 | P a g e