Islanding Detection in Distributed Generation
Islanding Detection in Distributed Generation
GENERATION
A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE DEGREE OF
Master of Technology
In
Electrical Engineering
By
TRUPTIMAYEE PUJHARI
Master of Technology
In
Electrical Engineering
By
TRUPTIMAYEE PUJHARI
Under the guidance of
Prof. B. D. SUBUDHI
Prof. S. R. SAMANTARAY
CERTIFICATE
This is to certify that the thesis entitled, Islanding Detection in Distributed
Date:
Prof. S.R.Samantaray
Co- Supervisor
ACKNOWLEDGEMENT
I would like to express my deep sense of profound gratitude to my honorable, esteemed guide,
Prof. B.D. Subudhi, Head of the Department electrical Engineering for his guidance and
constant support. Over the time he has introduced me to the academic world. His perspective on
my work has inspired me to go on. I am glad to work with him.
I would like to thank all my friends and especially PCD friends for all the thoughtful and mind
stimulating discussions we had, which prompted us to think beyond the obvious.
I cannot end without thanking my lovely family, on whose encouragement, support, and love, I
have relied through out my studies
Truptimayee pujhari
Roll No. 207EE215
Department of Electrical Engg
National Institute of Technology
Rourkela
CONTENTS
Chapter no
1
1.1
Title
Page No
INTRODUCTION
1.1.1
1.1.2
1.1.3
1.1.4
1.1.5
1.1.6
Generation
1.1.7
Islanding
1.1.8
10
1.1.9
11
Network
1.2
Objectives
12
1.3
12
1.4
Contributions
13
14
TECHNIQUES
2.1
2.2
14
2.1.1
15
2.1.2
17
18
2.2.1
18
2.2.2
20
2.2.3
23
26
ROCOF Relay
26
3.2
27
30
4.2
Simulation model
32
4.3
33
4.4
37
4.5
38
4.6
39
5
5.1
45
45
islanding detection
5.1.1 Islanding Detection based upon Negative
47
49
51
Islanding Situation
5.1.4
55
58
5.2.1
58
ii
5.2.2
64
65
6.1
Discussions
65
6.2
Conclusions
65
FUTURE SCOPE
66
REFERENCES
67
iii
ABSTRACT
The advancement in new technology like fuel cell, wind turbine, photo voltaic and new
innovation in power electronics, customer demands for better power quality and reliability are
forcing the power industry to shift for distributed generations. Hence distributed generation (DG)
has recently gained a lot of momentum in the power industry due to market deregulations and
environmental concerns. 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
distributed generators. An important requirement to interconnect a DG to power distributed
system is the capability of the DG to detect islanding detection. Failure to trip islanded
generators can lead to a number of problems to the generators and the connected loads. The
current industry practice is to disconnect all distributed generators immediately after the
occurrence of islands. Typically, a distributed generator should be disconnected within 100 to
300 ms after loss of main supply. To achieve such a goal, each distributed generator must be
equipped with an islanding detection device, which is also called anti islanding devices like
vector surge relay and ROCOF relay.
The contributions of this M-tech Thesis are as follows
Review of the current practices and research work developed on
distributed generation
Problems linked with distributed generations like islanding
Necessity of islanding detection
Survey on different islanding detection techniques
Study of wavelet transform
Development of new method for islanding detection of wind turbines
based on negative sequence component of voltage & current using the
wavelet transform. The detailed coefficient at the level-1 (d1) clearly
localizes the event and thus detects the islanding condition. Also the
change in energy and standard deviations of the detailed coefficients for
one cycle current and voltage signals distinguishes islanding conditions
from non-islanding ones.
iv
LIST OF FIGURES
1.1
1.2
1.3
1.4
10
2.1
14
2.2
15
2.3
16
2.4
17
2.5
22
3.1
26
Relay
3.2
27
Relay
3.3
29
29
4.1
32
4.2
35
MATLAB/SIMULINK model
36
4.4
38
4.5
39
4.6
43
5.1
47
Normal condition
5.2
47
vi
48
5.4
48
49
49
50
50
51
52
52
53
53
54
55
5.16
60
5.17
60
5.18
61
5.19
61
5.20
62
vii
5.21
62
Line cut-off
5.22
63
viii
64
LIST OF TABLES
1.1
1.2
2.1
25
5.1
56
5.3
5.4
5.5
5.6
ix
Chapter-1
Introduction
This thesis deals with a particular problem that occurs at the interface
between a distributed generation plant and the rest of the power system. The
problem can be described as islanding detection in power systems. The problem has
been investigated and discussed extensively in the last few years.
1.1
of distributed generator fed systems that has been tested on power distribution
busses of 25 kV and less. Recent interest in distributed generator installation into
low voltage busses near electrical consumers has created some new challenges for
protection engineers that are different from traditional radially based protection
methodologies. Therefore, typical protection configurations need to be re-thought
such as re-closures out-of-step monitoring, impedance relay protection zones with
the detection of unplanned islanding of distributed generator systems. The condition
of islanding, defined as when a section of the non utility generation system is
isolated from the main utility system, is often considered undesirable because of the
potential damage to existing equipment, utility liability concerns, reduction of
power reliability and power quality.
Current islanding detection methods typically monitor over/under voltage
and over/under frequency conditions passively and actively; however, each method
has an ideal sensitivity operating condition and a non-sensitive operating condition
with varying degrees of power quality corruption called the non detection zone
(NDZ). The islanding detection method developed in this thesis takes the
theoretically accurate concept of impedance measurement and extends it into the
symmetrical component impedance domain, using the existence of naturally and
artificially produced unbalanced conditions. Specific applications, where this
islanding detection method improves beyond existing islanding detection methods,
1.1.1
Human progress has been linked to the increase of energy consumed per
capita [1], [2]. In the last 20 years, electrical consumption has been steadily
increasing in North America at a rate of 1.1% for Canada, and 2.0% for the United
States [3]; however, the investment into new bulk electric power sources such as
hydro dams and nuclear generation plants has become politically, economically and
physically limited [4]. For example, transmission investment in the year 2000 was
$2.5 billion dollars less than the level of investment in 1975, where over this same
period, electricity sales nearly doubled [5]. At the current demand growth, the
United States bulk electric power system is estimated to be approximately 5 to 15
years away from the power demand exceeding the generation capacity as seen in
Figure 1.1. The United States has historically consumed a median of 7.5 times the
power of Canada which can be seen in the Canadian winter demand growth in
Figure 1.3.
Small localized power sources, commonly known as Distributed
Generation (DG), have become a popular alternative to bulk electric power
generation [6]. There are many reasons for the growing popularity of DG; however,
on top of DG tending to be more renewable, DG can serve as a cost effective
alternative to major system upgrades for peak shaving or enhancing load capacity
margins. Additionally, if the needed generation facilities could be constructed to
meet the growing demand, the entire distribution and transmission system would
also require upgrading to handle the additional loading. Therefore, constructing
additional power sources and upgrading the transmission system will take
significant cost and time, both of which may not be achievable. These trends are not
2
only limited to North America, but worldwide; the demand for electricity is
expected to double in the next 20 years [7].
The costs of power outages to a countrys economy can be staggering. The
cost associated with power outages to all business sectors in the United States has
been determined to be of the order $164 Billion US per year. More specifically, the
average cost of a power outage to a medium sized company is $1477 US for one
second and $7000 US for one hour. Though the cost of one second of outage is
considerable, the cost of one hour, which is a 3600 times longer duration is only 4.7
times of cost increase [8]; hence, initial quick outages are important to avoid
significant cost implications to the economy. Distributed Generators can assist in
reducing these occurrences by strengthening networks that are near to their stability
limit.
Figure 1.1: 2006 United States Projected Summer Generation and Capacity [13]
Typical
Utility interface
capacity
Photovoltaic
10VA
to
inverter
to
Induction
5000VA
Wind
10VA
500KVA
Geothermal
100VA
and
Synchronous
Generators, Inverters
to
Synchronous Generator
several MVA
Mycro Hydro
100VA
to
Induction
several MVA
Generator
Reciprocating
1000VA
to
Induction
Engine
several MVA
Generator
Combustion
1000VA
Synchronous Generator
Turbine
several MVA
Combined
1000VA
Cycle
several MVA
Micro
10 KVA to
Turbines
several MVA
Fuel Cells
10 KVA to
to
to
or
Synchronous
or
Synchronous
Synchronous Generator
inverter
inverter
several MVA
wind power, photo voltaic, hydro power. The figure 1.2 shows the single line
diagram of the distribution system with multiple DGs.
S/S
CB
PI
LINE
PI
LINE
PI
LINE
CB
CB
DG
DG
Figure 1.2: single line diagram of Distributed system with multiple DGs
1.1.5
1.1.6
Figure 1.3: 2006 Canadian Projected Winter Generation and Capacity [13]
where the findings of the various studies are listed in Table 1.2.
1.
2.
Voltage Flicker
3.
4.
5.
6.
7.
Transient Stability
8.
9.
10.
1.1.7
Islanding
Islanding is the situation in which a distribution system becomes electrically
isolated from the remainder of the power system, yet continues to be energized by DG
connected to it. As shown in the figure1.4. Traditionally, a distribution system doesnt
have any active power generating source in it and it doesnt get power in case of a fault in
transmission line upstream but with DG, this presumption is no longer valid. Current
practice is that almost all utilities require DG to be disconnected from the grid as soon as
possible in case of islanding. IEEE 929-1988 standard [24] requires the disconnection of
DG once it is islanded .Islanding can be intentional or Non intentional. During
maintenance service on the utility grid, the shut down of the utility grid may cause
islanding of generators. As the loss of the grid is voluntary the islanding is known. Nonintentional islanding, caused by accidental shut down of the grid is of more interest. As
there are various issues with unintentional islanding. IEEE 1547-2003 standard [25]
stipulates a maximum delay of 2 seconds for detection of an unintentional island and all
DGs ceasing to energize the distribution system,
1.1.8
Although there are some benefits of islanding operation there are some drawbacks
as well. Some of them are as follows:
Line worker safety can be threatened by DG sources feeding a system after
primary sources have been opened and tagged out.
The voltage and frequency may not be maintained within a standard
permissible level. Islanded system may be inadequately grounded by the DG
interconnection.
Instantaneous reclosing could result in out of phase reclosing of DG. As a
result of which large mechanical torques and currents are created that can damage
the generators or prime movers [26] Also, transients are created, which are
potentially damaging to utility and other customer equipment. Out of phase
reclosing, if occurs at a voltage peak, will generate a very severe capacitive
switching transient and in a lightly damped system, the crest over-voltage can
approach three times rated voltage .[27]
Various risks resulting from this include the degradation of the electric
components as a consequence of voltage& frequency drifts.
10
Due to these reasons, it is very important to detect the islanding quickly and
accurately.
1.1.9
islanding [28]. Their goal is to improve the distribution level (25 kV and below)
customer service reliability especially in regions where the reliability is below
customers needs. It is believed that customer reliability could improve with the
addition of DG sources and that the DG may be able to sell electricity back to the
utility. However, without complex studies and frequent expensive system upgrades
DG islanding is not allowed. Some examples of these studies are: real and reactive
power profile and control, planning for islanding, minimum/ maximum feeder
loading, islanding load profile, minimum/maximum voltage profile, protection
sensitivity and DG inertia. One more specific example is how substation autoreclosers of circuit breakers and main line reclosers may be disabled and other
protection devices may need to be removed to allow proper coordination of utility
sources and DG sources.
Maintenance times might also increase as utility workers will not only need
to lockout the utility lines but they will need to take additional time to lockout all
the installed DG lines. Some of the required installation studies of an Independent
power producer must complete to be able to island are: 1. inadvertent islanding and
planned islanding study, 2. reliability study, 3. power quality study, 4. utility
equipment upgrade assessment, 5. safety and protection reviews, and 6. commercial
benefit study. Clearly the costs of designing a DG to be capable of islanding or to
simply be installed into the main utility owned network requires extensive and
costly engineering and business reviews which may be outside the financial range
of smaller DG suppliers.
11
1.2 Objectives:
The objective of this thesis is to develop an islanding detection technique
which works effectively in Power distribution network with multiple DG interface.
The proposed research investigates the performance of the negative sequence
voltage components for islanding detection. As negative sequence components are
the key parameters for detecting unbalanced conditions in the power systems, the
same has been considered for islanding detection during the islanding process. The
Change in energy and standard deviations of the d1 Wavelet coefficients of the
negative sequence voltage and current detects the islanding conditions accurately.
Another important parameter such as negative sequence impedance has been
measured at the target DG location for islanding detection. Thus this thesis aims to
derive a new technique for islanding detection which has edge over the existing
conventional techniques.
1.3
12
1.4
Contributions:
Development of new method for islanding detection of wind turbines based
13
Chapter-2
Review of Islanding Detection Techniques
The main philosophy of detecting an islanding situation is to monitor the
DG output parameters and system parameters and/ and decide 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 2.1
14
This method has the advantages of its simplicity of control and its reliability.
In a radial system there is only one transmitting generator needed that can
continuously relay a message to many DGs in the network. The only times the
15
message is not received is if the interconnecting breaker has been opened, or if there
is a line fault that corrupts the transmitted signal.
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 have prohibitive cost barriers associated with it that may be
especially undesirable for the first DG system installed in the local network.
Another disadvantage is if the signaling method is applied in a non radial system,
resulting in the use of multiple signal generators. This scenario can be seen in
Figure 2.3 where the three feeder busses connect to one island bus. The
implementation of this system, opposed to a simple radial system, will be up to
three times the cost.
Figure 2.3: Distributed Generation Multi Power Line Signaling Islanding Detection
Issue
16
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 DGs, to either
remain in operation, or to discontinue operation. Transfer tip 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. This is
one of the most common schemes used for islanding detection [29].This can be seen
in figure 2.4
17
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.
Reconfiguration of this device in the planning stages of DG network is necessary in
order to consider if the network is expected to grow or if many DG installations are
planned. 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 and
special agreements may be necessary with the utility. If the transfer trip method is
implemented correctly in a simple network, there are no non-detection zones of
operation.
2.2
dp
, at
dt
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 load
change[ 30]. It has been found that this method is much more effective when
18
the distribution system with DG has unbalanced load rather than balanced load.
[31]
very high when the DG is islanded. The rate of change of frequency (ROCOF)
can be given by [32]
ROCOF,
Where,
df
dt
p
2 HG
Large systems have large H and G where as small systems have small H and
G giving larger value for df ROCOF relay monitors the voltage waveform and will
dt
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 DGs capacity matches with its local loads. However, an
advantage of this method along with he rate of change of power algorithm is that,
even they fail to operate when load matches DGs generation, any subsequent local
load change would generally lead to islanding being detected as a result of load and
generation mismatch in the islanded system.
with larger capacity. Rate of change of frequency over power utilize this concept to
determine islanding condition .Further more, test results have shown that for a small
19
power mismatch between the DG and local loads, rate of change of frequency over
power is much more sensitive than rate of frequency over time.[33]
20
(b)
m sin
f (k
2
fm
1)
fn
fn
(k1)
designed in such a way that its slope is greater than that of the phase of the load in
the unstable region. A SMS curve, with m =10 and fm = 53 Hz, is shown in Figure
21
2.5. When the utility is disconnected, operation will move through the unstable
region towards a stable operating point (denoted by black dots in Figure 2.5).
Islanding is detected when the inverter frequency exceeds the setting.
This detection scheme can be used in a system with more than one inverter
based DG. 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 [41].
22
2.2.3
The active technique is implemented only when the islanding is suspected by the
passive technique. Some of the hybrid techniques are discussed 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 [42] which is given as
VU=
V
Sq
v
v
sq
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.
Co-variance (Tav , Tv )
E Tav( n) U av Tv( n) U v
k
d
kd
Tv ( k )
Tav
Tv( k )
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. There is no single islanding detection
technique which will work satisfactorily for all systems under all situations. The
choice of the islanding detection technique will largely depend on the type of the
DG and system characteristics. Recently, hybrid detection techniques have been
proposed and it seems that the hybrid detection technique is the way to go with
passive technique detecting the islanding when change in system parameter is large
and initiating the active technique when the change in system parameter is not so
large for the passive technique to have an absolute discrimination.
Table 2.1 summarizes the islanding detection techniques, their advantage and
disadvantage, and examples.[44]
24
Advantages
Disadvantages
Examples
Expensive to implement
especially for small
systems.
Techniques
Remote
Highly reliable
Techniques
2
Local
Techniques
a. Passive Techniques
Short
detection
time
Do not perturb the
system
Accurate
when
there is a large
mismatch
in
generation
and
demand in the
islanded system.
Difficult to detect
islanding when the
load and generation in
the islanded system
closely match
Special care has to be
taken while setting the
thresholds
If the setting is too
aggressive then it
could
result
in
nuisance tripping
b. Active techniques
Can
detect islanding
even in a perfect
match
between
generation
and
demand in the
islanded system
(Small NDZ
Introduce perturbation
in the system
Detection time is slow
as a result of extra
time needed to see the
system response for
perturbation
Perturbation
often
degrades the power
quantity
and
if
significant enough, it
may degrade the
system stability even
when connected to
the grid
c .Hybrid Techniques
Have
small NDZ.
Perturba
tion is introduced
only
when
islanding
is
suspected
Islanding
detection
time is prolonged as
both
passive
and
active technique is
implemented
25
Rate of change of
output power scheme
Rate of change of
frequency scheme
Rate of change of
frequency over power
scheme
Change of impedance
scheme
Voltage
unbalance
scheme
Harmonic
distortion scheme
Reactive
power
export
error
detection scheme
Impedance
measurement
scheme
Phase (or frequency)
shift schemes (like
SMS, AFD, AFDPF
and ALPS
Technique based on
positive feedback and
voltage imbalance
Technique based on
voltage and reactive
power shift
Chapter- 3
Existing Voltage Surge and ROCOF relay for Islanding
Detection
3.1 ROCOF Relay
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.
27
larger load (or smaller) because the current ISYS provided (or consumed) by the
power grid is abruptly interrupted. Thus, the generator begins to decelerate (or
accelerate).
Consequently, the angular difference between VT and EI is suddenly
increased (or decreased) and the terminal voltage phasor changes its direction, as
shown in Fig. 3.3(b). Analyzing such phenomenon in the time domain, the
instantaneous value of the terminal voltage jumps to another value and the phase
changes the phase position changes as depicted in Fig. 3.4, 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 surge. 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. 3.4. 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.
To avoid false operation, ROCOF and VS relays are disabled if the terminal
voltage decreases below a determined voltage threshold. Showed that ROCOF
relays require a smaller active power imbalance level than VS relays for successful
islanding detection. On the other hand, ROCOF relays are much more susceptible to
false operation than VS relays.
28
Figure 3.3: Internal and terminal voltage phasors (a) before opening with CB (b)
after opening with CB
29
Chapter-4
Proposed Techniques and System Studied
4.1
30
The negative sequence component of the voltage and current signals are
extracted from the derived voltage and current signal at the target DG locations.
The one cycle negative sequence voltage and current signal are processed through
Wavelet transform (db4). The time-frequency information derived at the level-1
decompositions (d1) localizes the corresponding islanding events. Further to
provide a threshold for detecting islanding conditions from non-islanding ones, the
standard deviations (std) and change in energy (ce) of the d1 level coefficients for
one cycle are computed.
31
Figure 4.1: The studied Power Distribution network with multiple DGs
32
Transformer T2, T3, T4 and T5: rated MVA = 10, f = 50 Hz, rated kV = 575 V/
25 kV, Vbase = 25 kV, R1 = 0.00375 pu, X1= 0.1 pu, Rm= 500 pu, Xm= 500 pu
Distribution lines (DL): DL-1, DL-2, DL-3 and DL-4: PI-Section, 30 km each,
Rated kV = 25, rated MVA = 20, Vbase = 25 kV, R0 = 0.1153 ohms/km, R1 = 0.413
ohms/km, L0 = 1.05e-3 H/km, L1 = 3.32e-3 H/km, C0 = 11.33e-009 F/km, X1 =
5.01e-009 F/km,
Normal Loading data: L1 = 15 MW, 5 MVAR. L2, L-3, L4, L5=8.0MW, 3
MVAR.
The voltage and current signals are retrieved at the target DG location for
islanding conditions and non-islanding conditions (other disturbances). The relays
for each DG units are placed at the DG end. For example, the relay for CB_DG-1 is
placed at DG-1 to collect the voltage and current information for both islanding and
non-islanding conditions. The possible situations of islanding and non-islanding
conditions studied are given as follows
on 50 Hz base frequency). The voltage and current signals are retrieved at the target
DG location (DG-1, DG-4). The islanding starts at 0.3 sec as shown in the Fig.4.2.
The complete simulation is carried out using MATLAB-SIMULINK software
package.
34
voltage(pu)
0.5
0
-0.5
current(pu)
-1
0.2
0.25
0.3
0.35
0.4
0.45
0.5
0.35
0.4
time in second
0.45
0.5
1
0
-1
0.2
0.25
0.3
Figure 4.2: Three-Phase Voltage and Current Signals under Islanding Condition
retrieved at the target DG location (starts at 0.3 sec)
35
10
Wind (m/s)
<Vdc>
[Vdc]
aA
bB
cC
c
cC
C
C
C
2500 MVA B120 120 kV/25 kV Three-Phase Breaker B25
X0/X1=3 (120 kV) 47 MVA
(25 kV)
120 kV
a
b
c
aA
aA
bB
bB
20 km line
A
B
C
N
<P>
cC
B575
(575 V)
25 kV/ 575 V
6*2 MVA
Load
15MW
[wr]
<wr>
m
A
B
C
A
B
C
A a
B b
C c
<Q>
[P_pu]
[Q_pu]
Wind Turbine
Doubly -Fed Induction Generator
Filter
(Detailed Model )
0.9 Mvar
Q=50
9 MW Wind Farm
(6 x 1.5 MW)
Grounding
Transformer
X0=4.7 Ohms
3.3ohms
10
Wind (m/s)
<Vdc>
[Vdc_1]
aA
bB
cC
20 km line 1
Load
15MW1
B575
(575 V)1
25 kV/ 575 V
6*2 MVA1
<wr>
<P>
A
B
C
Three-Phase Breaker 1
A
B
A
B
C
A a
B b
C c
<Q>
[wr_1]
[P_pu_1]
[Q_pu_1]
Wind Turbine
Doubly -Fed Induction Generator
Filter
(Detailed Model )1
0.9 Mvar
Q=1
9 MW Wind Farm
(6 x 1.5 MW)
Vabc_B25
signal THD
THD Va _B25
Discrete
Total Harmonic
Distorsion
vol .mat
The model parameters , initial conditions and sample times
(Ts_Power =5e-6 s and Ts_Control =100 e-6 s)
are automatically initialized (see Model Properties )
seqvol .mat
To File
cur.mat
To File 3
seqcur.mat
To File 4
To File 1
Discrete,
Ts= 5e-006 s.
Vabc_B575
powergui
Iabc _B575
seqvolph .mat
Vabc_B575 (pu)
To File 2
Iabc_B575 (pu)
[P_pu]
seqcurph .mat
To File 5
P (MW)
-K[Q_pu]
Q (Mvar )
MW
[Vdc]
[wr]
Vabc_B576
Iabc _B576
Vdc (V)
wr (pu)
Vabc_B576 (pu)
Iabc_B576 (pu)
Scope
Mag
abcPhase
Discrete 3-phase
Sequence Analyzer
Mag
abcPhase
Discrete 3-phase
Sequence Analyzer 1
36
Scope 1
4.4
Condition-1:
Both the three phase circuit breaker& three phase circuit breaker1 are in
closed conditions. That is the normal condition .So the voltage& current waveforms
found at any DG end is purely sinusoidal.
Condition-2:
Three phase circuit breaker is initially closed and is opened with
transition time.3 to .5 . Three phase circuit breaker 1 is permanently closed. As after
switching off the three phase breaker whole part of the distribution system is
isolated from the remainder part of the power system. So this is the islanding
condition.
Condition-3:
Three phase circuit breaker is permanently closed. Three phase breaker1
is initially closed and is opened with transition time .3 to .5. As another DG is there
in the line where three phase breaker1 is connected. This is the DG line trip
condition.
Condition-4:
This is the sudden load change condition. Where suddenly load is
changed up to 50%.
For the above four events experimental study is observed and islanding is
detected by developing some algorithm.
37
4.5
There are some new techniques we have developed for islanding detection.
1 phase Voltage and current waveforms observed at the DG-1 location during
islanding:
1
0.8
0.6
0.4
voltage
0.2
0
-0.2
-0.4
-0.6
-0.8
-1
50
100
150
200
sample(time)
250
300
350
38
400
0.5
current
-0.5
-1
-1.5
50
100
150
200
sample(time)
250
300
350
400
4.6
Wavelet Transform
Given a function f(t), its continuous Wavelet Transform (CWT) can be
calculated as follows:
CWT( f , x, y )
f (t ) *
a
40
y
x
dt
(4.1)
DWT (f, x, y)
f (k ) *
n kx0m
(4.2)
x0m
m
Where the parameters x and y in (5.1) are replaced by x0 and kx0 , k and
m being integer variables. In a standard DWT, the coefficients are sampled from the
CWT on a dyadic grid.
Associated with the wavelet is a scaling function
along with the wavelet function creates a multi-resolution analysis (MRA) of the
signal. The scaling function of one level can be represented as a sum of a scaling
function of the next finer level.
(t )
h(n) 2 (2t
n)
(4.3)
(t )
h1 ( n) 2 (2t
n)
(4.4)
Where
h(k)
( 1) k h(1 k )
h1 (k )
(4.5)
y(t )
c jo (k )2
jo
d j (k )2 2 (2 j t k )
(2 t k )
41
j jo
(4.6)
c j (k )
c j 1 ( m) h ( m 2 k )
(4.7)
d j (k )
c j 1 ( m) h1 ( m
2k )
(4.8)
Equations (4.7) and (4.8) show that the coefficients at a coarser level can be
attained by passing the coefficients at the finer level to their respective filters
followed by a decimation of two. This will result in the number of samples in the
coarser level to be approximately half of the number of samples at the finer level.
For a signal that is sampled at a frequency higher than the Nyquist
frequency, the samples are used as c j
(m) .
Bk
fk
fs
2k 1
3 fs
2k 2
42
(4.9)
(4.10)
low pass filters at each scaling stage of Wavelet Transform. This can be thought of
as successive approximations of the same function, each approximation providing
the incremental information related to a particular scale (frequency range), the first
scale covering a broad frequency range at the high frequency end of the frequency
spectrum, however, with progressively shorter bandwidths. Conversely, the first
scale will have the highest time resolution; higher scales will cover increasingly
longer time intervals. While in principle any admissible wavelet can be used in the
Wavelet analysis, Daubechies Wavelet (DB4) is used in this work for both the
purposes of fault section identification and phase selection for a transmission line.
If the used scaling function and the Wavelet function form an orthogonal
basis, then Parsevals theorem relates the energy of the distorted signal to the
energy in each expansion coefficients and their Wavelet coefficients. This means
that the norm of energy of the signal can be partitioned in terms of expansion
coefficients .This feature is utilized here to differentiate different faults. The energy
of the distorted signal will be partitioned at different resolution levels in different
ways depending on the signals to be analyzed. The Energy of the signal is given by
43
Esignal
y (t ) dt
c(k )
k
d j (k )
2
(4.11)
j jo k
The change in energy is found out by deducting the energy content of the d1
coefficients for one cycle signal before islanding inception from the energy content
of the d1 coefficients for one cycle signal after islanding inception. Similarly the
standard deviation of the d1 coefficients for one cycle signal is computed for
detecting islanding events from non-islanding ones .Standard deviation can be
considered as a measure of the energy for a distorted signal with zero mean and is
utilized in this work as a feature to detect the islanding from non islanding one, in
power distribution network with multiple DG interface.
44
Chapter 5
Simulation Results & Discussion
The test system is simulated with sampling rate 1.6 KHz. Sampling period is
5e-006.Samples/cycle is 32.The four events described above are simulated. Here the
steady state operation of the DFIG and its dynamic response for a remote fault on
the 120-KV system is observed. The voltage and current wave forms are observed
at DG terminals by closing and opening of the two circuit breakers.
Vn
1
(Va a 2Vb aVc )
3
In
1
( I a a 2 I b aI c )
3
currents retrieved at the taget DG location, and a 1 120 , is the complex operator.
The negative sequence components are found out using sequence analyser block of
Simulink.
The negative sequence voltage and currents are processed through Wavlet
Tranfrom (db4) for time localization of the islanding event. The negative sequence
voltage and corresponding d1 coefficints for different islanding and nonislanding
conditions are shown in Fig. 5.1,5.2,5.3,5.4. The negative sequence current and
corresponding d1 coefficients for different islanding and nonislanding conditions
are shown in Fig. 5.5,5.6,5.7,5.8. The d1 coefficients clearly localizes the islanding
event and thus helps in detecting the same.
46
5.1.1
1.2
x 10
1
0.8
0.6
0.4
0.2
0
50
100
150
200
sample(time)
250
300
350
400
-4
x 10
d1
2
1
0
-1
-2
20
40
60
80
100
sample(time)
120
140
160
180
200
Figure 5.1: The negative sequence component of voltage and d-1 coefficient for
Normal condition
Condition-2 (Islanding Condition)
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
50
100
150
200
sample(time)
250
300
350
400
-3
x 10
4
2
d1
0
-2
-4
-6
-8
20
40
60
80
100
sample(time)
120
140
160
180
200
Figure 5.2: The negative sequence component of voltage and d-1 coefficient for
Islanding condition
47
x 10
4
3
2
1
0
50
100
150
200
sample(time)
250
300
350
400
-4
x 10
6
4
d1
2
0
-2
-4
-6
20
40
60
80
100
sample(time)
120
140
160
180
200
Figure 5.3: The negative sequence component of voltage and d-1 coefficient for
DG line trip condition
0.025
0.02
0.015
0.01
0.005
0
50
100
150
200
sample(time)
250
300
350
400
-3
x 10
d1
0
-2
-4
-6
20
40
60
80
100
sample(time)
120
140
160
180
200
Figure 5.4: The negative sequence component of voltage and d-1 coefficient for
sudden load change condition
48
x 10
5
4
3
2
1
0
50
100
150
200
sample(time)
250
300
350
400
-4
x 10
d1
2
0
-2
-4
-6
20
40
60
80
100
sample(time)
120
140
160
180
Figure 5.5: The negative sequence component of current and d-1 coefficient for
normal condition
Islanding Condition
0.14
0.12
0.1
0.08
0.06
0.04
0.02
0
50
100
150
200
sample(time)
250
300
350
400
0.015
0.01
d1
0.005
0
-0.005
-0.01
-0.015
20
40
60
80
100
sample(time)
120
140
160
180
200
Figure 5.6: The negative sequence component of current and d-1 coefficient for
Islanding condition
49
200
0.01
0.008
0.006
0.004
0.002
0
50
100
150
200
sample(time)
250
300
350
400
-4
x 10
4
2
d1
0
-2
-4
-6
-8
20
40
60
80
100
sample(time)
120
140
160
180
Figure 5.7: The negative sequence component of current and d-1 coefficient for
DG line trip condition
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
50
100
150
200
sample(time)
250
300
350
400
-3
x 10
4
2
d1
0
-2
-4
-6
-8
20
40
60
80
100
sample(time)
120
140
160
180
200
Figure 5.8: The negative sequence component of current and d-1 coefficient for
sudden load change condition
50
200
-3
x 10
islanding
normal
4
-2
-4
-6
-8
20
40
60
80
100
sample(time)
120
140
160
180
200
Figure 5.9: Comparison between d-1 coefficient of negative sequence voltage for
islanding and normal condition
51
-3
x 10
islanding
load change
4
-2
-4
-6
-8
20
40
60
80
100
sample(time)
120
140
160
180
200
Figure 5.10: Comparison between d-1 coefficient of negative sequence voltage for
islanding and load change (up to 50%) condition
-3
x 10
islanding
DG line cutoff
4
-2
-4
-6
-8
20
40
60
80
100
sample(time)
120
140
160
180
200
Figure 5.11: Comparison between d-1 coefficient of negative sequence voltage for
islanding and DG line cutoff condition
52
Similar studies are made for different islanding and non islanding condition based
upon the d1 coefficients of negative sequence currents. The results are described in
fig. 5.12, 5.13, 5.14. Similar conclusions are drawn.
0.015
islanding
load change
0.01
0.005
-0.005
-0.01
-0.015
20
40
60
80
100
sample(time)
120
140
160
180
200
Figure 5.12: Comparison between d-1 coefficient of negative sequence current for
islanding and Load change condition
0.015
islanding
DG line cutoff
0.01
0.005
-0.005
-0.01
-0.015
20
40
60
80
100
sample(time)
120
140
160
180
200
Figure 5.13: Comparison between d-1 coefficient of negative sequence current for
islanding and DG line cutoff condition
53
0.015
Islanding
Normal
0.01
0.005
-0.005
-0.01
-0.015
20
40
60
80
100
sample(time)
120
140
160
180
200
Figure 5.14: Comparison between d-1 coefficient of negative sequence current for
islanding and normal condition
54
5.1.4
Acquisition of negative
sequence voltage
d1 coefficients of
sequence voltage
negative
STD>VTH
No
Non Islanding
yes
Islanding
End
Figure 5.15: Proposed algorithm for negative sequence of voltage relay
Events
Change in
energy (E)
1-Islanding
condition
2-Normal
condition
.1225
Change
in
standard
deviation(std)
.0234
.0041
6.9840e-004
3-DG
line
cutoff
condition
4-sudden
load change
5.4797e-004
8.5193e-005
.00675
.00121
56
Change
in
energy((E)
1-Islanding
condition
2-Normal
condition
3-DG line
cutoff
condition
4-sudden
load change
.1153
Change
standard
deviation(std)
0.0264
0.0065
0.0019
0.0013
0.0029
0.0032
0.0085
in
Change
in
energy((E)
1-Islanding
condition
2-Normal
condition
3-DG line
cutoff
condition
4-sudden
load change
0.1315
Change
in
standard
deviation(std)
.0127
0.0021
5.24550e-004
6.3167e-004
7.8934e-005
0.00524
0.00142
Change
in
energy((E)
1-Islanding
condition
0.1256
Change
in
standard
deviation(std)
0.0189
2-Normal
condition
3-DG line
cutoff
condition
4-sudden
load change
0.0087
0.0021
0.0013
0.0018
0.0032
0.0098
57
Zn
Vn
In
where Vn is the negative sequence voltage and I n is the negative sequence current
derived at taget DG location.
5.2.1 Negative sequence impedance for different Islanding and Non Islanding
Situations
The negative sequence imepdance for different islanding and nonislanding
conditions such as nomal,islanding, DG line trip and sudden load change
conditions are shown in Fig. 5.16,5.17,5.18,5.19 respectively. Similarly, negative
sequence impedance comparision between islanding and non-islanding conditions
such as normal operation, sudden load change (non-islanding) and DG line cut-off
are shown in Fig. 5.20, Fig. 5.21and 5.22 respectively. It is found that the negative
sequence impedance becomes steady after islanding compared to non-islanding
situations.
To further provide a thershold for islanding detection, the standard
deviations of the negative sequence imepdance for one cycle is found out and given
in Table- 5.5 and 5.6. It is found that the standard deviation is 0.0152 compraed to
0.3551,0.2977,0.1336 for non-islanding cases at target DG location DG-1. To
furhter know the efect of changing DG locations, the similar observations are made
for standard devtaions for negative equence impadance at target DG location DG-2.
58
It is seen that the standard deviation is very low for islanding condition
compared to non-islanding ones, and thus providing a threshold effectively
distinguishes the islanding events from non-islanindg conditions. It is observed that
the negative sequence impedance is marginally affected when the target DG
location is changed. Thus the negative sequence impedance is a potential measure
for detecting islanding conditions in distributed generations.
Change in standard
deviation((std)
Islanding
condition
Normal
condition
DG
line
cutoff
Sudden load
change
0.0152
0.3521
0.2977
0.1336
Conditions
Change in standard
deviation((std)
0.0172
Islanding
condition
Normal condition
0.4123
DG line cutoff
0.3178
Sudden
change
0.1542
load
59
Condition-1
2.5
1.5
0.5
50
100
150
200
sample(time)
250
300
350
400
Condition-2
3
2.5
1.5
0.5
50
100
150
200
sample(time)
250
300
350
60
400
Condition-3
2.5
1.5
0.5
50
100
150
200
sample(time)
250
300
350
400
Figure 5.18: The negative sequence Impedance for DG line trip Condition
Condition-4
2.5
1.5
0.5
50
100
150
200
sample(time)
250
300
350
Figure 5.19: The negative sequence Impedance for sudden load change Condition
61
400
1.5
0.5
50
100
150
200
samples(time)
250
300
350
400
3
islanding
DG line cutoff
2.5
1.5
0.5
50
100
150
200
sample(time)
250
300
350
400
62
3
islanding
normal
2.5
1.5
0.5
50
100
150
200
sample(time)
250
300
350
400
Figure 5.22: The negative sequence impedance comparison between islanding and non
islanding condition (normal condition)
63
5.2.3 Proposed Algorithm for islanding detection based upon Negative Sequence
Impedance
Start
Acquisition of negative
sequence voltage and current
Determination of Negative
Sequence impedance
Calculation of Standard
deviation of impedance ZSTD
ZSTD< ZTH ?
No
Non islanding
Yes
Islanding
End
Figure 5.23: Negative sequence impedance based relay for islanding detection
64
Chapter-6
Discussions & Conclusions
6.1
Discussions
The proposed research investigates the potential of negative sequence
6.2
Conclusions
coefficients for one cycle signal data is found out which clearly detects the
islanding conditions. Also the negative sequence impedance is found out for both
islanding and non-islanding events, and it is observed that the standard deviation of
the negative sequence impedance of islanding event is very low compared to nonislanding condition, thus able to detect the islanding events effectively. Thus the
proposed methods are highly effective for islanding.
Future Scope:
More studies are required to meet the problems caused due to sudden load
change which may create false alarm in the islanding detection process.
66
REFERENCES
[1]
Jeremy Rifkin. The Hydrogen Economy: The Creation of the Worldwide Energy Web
and the Redistribution of Power on Earth, volume 1. Penguin Putnam Inc., New York,
NY, 1 edition, 2002.
[2] Steve Gehl. Electricity technology roadmap: 2003 summary and synthesis. Technical
report, Electric Power Research Institute, November 2003.
[3] Kevin J. Dasso. The reliability of bulk electric systems in North America. Technical
report, North American Electric Reliability Council, September 2006.
[4] John Micklethwait. Atomic renaissance. The Economist, 384(8545):102104,
September 2007.
[5] Edison Electric Institute. Eei survey of transmission investment. Report, May
2005.
[6] Manisa Pipattanasomporn, Michael Willingham, and Saifur Rahma. Implications of
on-site distributed
Generation for commercial/industrial facilities. IEEE Transactions of Power Systems,
20(4):206212, 2005.
[7] Mossadiq Umedaly. A vision for growing a world-class power technology cluster in a
smart, sustainable british columbia. British Columbia Reports and Publications, 2005.
[8] David Lineweber and Shawn McNulty. The cost of power disturbances to industrial
and digital
Economy companies. Technical report, Primen, June 2001.
[9] Spencer Abraham. National transmission grid study. Technical report, US Department
of Energy,
May 2002.
[10] S. Knoke, J.Majeski, K. Smit, and K. Tarp. Installation, operation, and
maintenance costs for distributed
Generation technologies. Technical Report 1007675, Electric Power Research Institute,
February 2003.
[11] Electrical Power Research Institute. 1999 summary and synthesis. Technical Report
CI-112677- V1, Electrical Power Research Institute, July 1999.
67
[24] Recommended Practice for Utility Interconnected Photovoltaic (PV) Systems, IEEE
Standard 929-2000, 2000.
[25] IEEE Standard for Interconnecting Distributed Resources into Electric
Power Systems, IEEE Standard 1547TM, June 2003.
[26] R. A. Walling, and N. W. Miller, Distributed generation islanding implications on
power system dynamic performance, IEEE Power Engineering Society Summer
Meeting, vol.1, pp. 92-96, 2002.
[27] A. Greenwood, Electrical Transients in Power Systems, New York: Wiley, 1971, pp.
83.
[28] BC Hydro Systems Group. Distribution power generator islanding guidelines.
BC Hydro, 2006.
[29] Ward Bower and Michael Ropp. Evaluation of islanding detection methods for
photovoltaic utility-interactive power systems. Report IEA PVPS Task 5 IEA PVPS T509: 2002, Sandia National Laboratories Photovoltaic Systems Research and
Development, March 2002.
[30] 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, pp. 948-954, July 1993
[31] 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, pp. 1249-1254, July
1995.
[32] 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, pp. 4.3.1-12.
[33] 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, pp. 346-351, 2001.
[34] 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, pp. 571-574, April 2004.
69
70
[48] A.H. Osman, O.P. Malik, Transmission line protection based on wavelet transforms,
IEEE Trans. Power Deliv. 19 (2) (2004) 515523.
[49] D. Chanda, N.K. Kishore, A.K. Sinha, A wavelet multi-resolution analysis for
location of faults on transmission lines, Electr. Power Energy Syst. 25 (2003) 5969.
[50] Cheng-Tao Hsieh, Jeu-Min Lin and Shyh-Jier Huang Enhancement of
islanding-detection of distributed generation systems via wavelet transform-based
approaches International Journal of Electrical Power & Energy Systems Volume 30,
Issue 10, December 2008, Pages 575-580.
71