Professional Documents
Culture Documents
May 2008
To my beloved mother and father
ACKNOWLEDGEMENTS
Particularly thanks to God for the blessing that gives me a patience and
courage in finishing my project report. Firstly, I would like to take this
opportunity to thank Prof. Madya Dr. Zulkurnain Bin Abdul Malek, my thesis
supervisor for his unfailing enthusiasm and encouragement. I also would like to
thank him for all his valuable advice and assistance, suggestions and comments
in completing this project report.
DECLARATION ii
DEDICATION iii
ACKNOWLEDGEMENT iv
ABSTRACT v
ABSTRAK vi
TABLE OF CONTENTS vii
LIST OF TABLES xi
LIST OF FIGURES xii
LIST OF ABBREVIATIONS xv
LIST OF SYMBOLS xvi
1 INTRODUCTION 1
1.1 Introduction 1
1.2 Objectives 2
1.3 Scope of Work 2
1.4 Project Flow Chart 3
1.5 Organisation of Thesis 4
2 LITERATURE REVIEW 5
2.1 Introduction 5
2.2 Case Study by Zia Emin and Yu Kwong Tong: 5
Ferroresonance Experience in UK: Simulation and
Measurement
2.2.1 Objective 5
2.2.2 Voltage Transformer 6
2.2.2.1 Comparison Field Measurement and Simulation 7
Result
2.2.3 Single Phase Traction Supply Transformer 8
2.2.3.1 Comparison Field Measurement and Simulation 9
Result
2.2.4 Conclusion 9
2.3 Case Study by YK Tong: NGC Experience on Ferroresonance 10
In Power Transformer and Voltage on HV Transmission
System
2.3.1 Introduction 10
2.3.2 Power Transformers 10
2.3.3 Voltage Transformer 11
2.3.4 Measurement to Predict or Prevent Ferroresonance 11
2.4 Case Study by David A.N. Jacobson: Example of 12
Ferroresonance in High Voltage Power System
2.4.1 Objective 12
2.4.2 Wound Potential Transformer-Circuit Breaker 12
Grading Capacitor
2.4.2.1 Description of Disturbance 12
2.4.2.2 Simulation Result 13
2.4.2.3 Mitigation Options 13
2.4.3 Transformer – Circuit Breaker Grading Capacitor 14
2.4.3.1 Description of Disturbance 14
2.4.3.2 Simulation Result 14
2.4.3.3 Mitigation Options 15
2.4.4 Open Delta Potential Transformer 15
2.4.4.1 Description of Disturbance 16
2.4.4.2 Simulation Result 16
2.4.4.3 Mitigation Options 17
2.4.5 Conclusion 17
2.4 Summary 17
3 FERRORESONANCE 19
4 METHDOLOGY 32
4.1 System Modeling 32
4.2 ATP-EMTP Simulation 32
4.3 Selected model and Validation 33
4.4 Resistor and Capacitor Model 34
4.5 Overhead Transmission Lines 34
4.6 Transformer Model 34
4.6.1 Nonlinear and frequency Dependent Parameter 36
4.6.1.1 Modelling of Iron Core 36
4.6.1.2 Modelling of Eddy Currents Effects 37
REFERENCES 72
LIST OF TABLES
DC - Direct Current
AC - Alternating Current
ATP - Alternative Transient Program
EMTP - Electro Magnetic Transient Program
PT - Potential Transformer
GB - Grading Bank
HV - High Voltage
LV - Low Voltage
LIST OF SYMBOLS
XL - Inductance reactance
XC - Inductance capacitance
C - Capacitance
L - Inductance
C - Capacitive
R - Resistance
VL - Load Voltage
E - Voltage source
f - Frequency
w - Frequency
Z - Impedance
CHAPTER 1
INTRODUCTION
1.1 Introduction
1.2 Objective
To solve the problem, one has to first study the problem and come up with
the process flow chart, which will guide the simulator throughout the project. It is
also to give the simulator the basic overview of the system and what is required
before simulations be completed. The flow chart of this project is as shown in Figure
1.1.
Review Ferroresonance
simulation done in power
system
Recommendation and
Conclusion
LITERATURE REVIEW
2.1 Introduction
2.2.1 Objective
The system arrangement in Figure 2.1 can be reduced to equivalent circuit in Figure
2.2.
When the same system is tested for a second time by changing the parameter
the response is totally different from earlier and it become into sub-harmonic
ferroresonance. The shapes of measured and simulated waveforms in Figure 2.3 and
2.4 are a good match.
Figure 2.3 shows the single line diagram of a traction supply transformer
arrangement. The single phase transformer is fed from two phases of the three-phase
system. Ferroresonance happens upon opening of circuit breakers 1 and 2 to de-
energize the line and the transformer. Because energy is coupled to the de-energized
network from the adjacent live parallel circuit through the inter-circuit coupling
capacitance, C series in the equivalent ferroresonant circuit. The equivalent C shunt in
the ferroresonance circuit is the phase-to-earth capacitance of the line and
transformer winding and bushing capacitance. The single line diagram in figure 2.5
can reduced to equivalent circuit in Figure 2.2.
Figure 2.6 and 2.7 shows the simulation result of field measurement and
digital simulation of a traction supply transformer. Obviously, the two phases
generate the transformer in prolong fundamental frequency mode ferroresonance.
2.2.4 Conclusion
It can be concluded that the ATP model used in this case study is capable of
predicting feasible ferroresonance modes as long as the parameters of the system are
known or can be estimated.
2.3 Case Study by YK Tong: NGC Experience on Ferro resonance in Power
Transformer and Voltage Transformer on HV Transmission Systems [2]
2.3.1 Introduction
2.4.1 Objective
Figure 2.8 demonstrates the single line diagram of wound power transformer
circuit at Dorsey s HVdc converter station 230kV. The station consists of four bus
sections on which the converter valves and transmission lines are terminated. Bus
A2 was removed for replacement breakers and to perform disconnects maintenance
and trip testing. Afterwards, potential transformer (V13F) failed. The switching
procedure de-energized bus and power transformer connected to energize B2
through grading capacitor. Normally connected to bus A2, but had been
disconnected. The ferroresonance effects failure to the power transformer.
Figure 2.8: Single line diagram of wound power transformer circuit
The mitigation option of this case is being connected 200 ohm loading
resistors at secondary service transformer SST1 and SST2.
Figure 2.12 shows the single diagram of open –delta potential transformer.
The purpose of grounding bank is to provide zero sequence current during single
line-to-ground for ground fault protection relay to operate. Generally, the open –
delta potential transformer as a backup to ground bank to taken out for service for
maintenance.
2.4.5 Conclusion
2.5 Summary
The first case study, it has been proven that ferroresonance in voltage
transformers and power transformers occurs due to nonlinear and dynamics in the
system. The case study analyses the performance of ferroresonance in the system by
comparing measurement data and the ATP-EMTP simulation results. The study
shows the measured and simulated voltages agree very well.
The second case study describes the ferroresonance phenomena when
connecting power transformers with overhead lines, or when connecting voltage
transformers with isolated sections of busbars. The energy in the capacitance of the
parallel lines or open circuit breaker grading capacitance is responsible for the
ferroresonance occurrence at 50Hz fundamental frequency or sub-harmonics
frequencies at 331/2Hz, 162/3Hz and 10Hz. The mitigation methods of
ferroresonance are by disconnecting the power transformer from the feeder, and for
the case voltage transformers, by adding damping devices at the secondary side of
the voltage transformer.
FERRORESONANCE
3.1 Introduction
Ferroresonance does not occur regularly and it is very hard to analyze. Any
response to a sudden change in the system may jump out of a steady state into
ferroresonance condition. Therefore, the basics of ferroresonance must be
identifying. The simple RLC circuit in Figure 3.1 can be used as an aid in explaining
resonance and hence ferroresonance.
Figure 3.1: Resonance in RLC Circuit
This linear circuit is resonating when at some given source of frequency; the
inductive reactance (XL) and capacitive reactance (XC) cancel each other out. These
impedance values be forecast and they change with the frequency. Capacitance (C)
will always have a capacitive reactance as in equation (1) and inductance (L) will
always have an inductive reactance as in equation (1), where w is the frequency of
the source.
XC = 1/ jwC (1)
XL = jwL (2)
I = V / (R + XL - XC ) *V / R (3)
The current in equation (3), depends on the resistance (R). If this resistance is
small, then the current can become very large in the RLC circuit. The size of current
during resonance is equation (3).
If the inductor in Figure (2) replaced by an iron cored non-linear
inductor, the exact values of voltage and current cannot be predicted as in a
linear model. Equation (3) will not indicate the size of current produced. The
inductance becomes nonlinear due to saturation of flux in the iron core. The
understanding that ferromagnetic material saturates is very important.
Ferromagnetic material has a property of causing an increase to the magnetic
flux density, and therefore magnetic induction.
As the current is increased, the magnetic flux density until a certain point
where the slope no longer linear and an increase in current lead to smaller and
smaller increases in magnetic flux density. This has been identified as the
saturation point. Figure 3.2 illustrates the relationship between magnetic flux
density and current. As the current increases in a ferromagnetic coil past the
saturation point, the inductance of the coil changes very quickly. This allows the
current to take on very unsafe high values. These high currents make damage on
the transformer. Most transformers have cores made from ferromagnetic
material. This is why ferroresonance is a concern for transformer operation.
Figure 3.3 is a basic series ferroresonance circuit and the curve obtained
from of the circuit shown in figure 3.5 and 3.6. These curves, illustrate the
following characteristics of ferroresonance:
Figure 3.4 illustrates the peak voltage VL at the terminals of the non-
linear inductance as a function peak amplitude E of the sinusoidal voltage
source.
Figure 3.4: Sensitivity to system parameters and the jump phenomenon
i. Fundamental mode
ii. Subharmonic mode
iii. Quasi –periodic mode
iv. Chaotic mode
Figure 3.7 shows the diagrams to explain the sub harmonic modes. The
signals are periodic with a period nT which is a multiple of the source period.
This state known as sub harmonic n or harmonic 1/n. Sub harmonic
ferroresonant states are normally of odd order. The spectrum presents a
fundamental equal to f0/n (where f0 is the source frequency and n is an integer)
and its harmonics (frequency f0 is thus part of the spectrum)
Figure 3.8 shows the diagram for the quasi-periodic mode. This mode is
not periodic. The spectrum is a discontinuous spectrum whose frequencies are
expressed in the form: nf1+mf2 (where n and m are integers and f1/f2 an irrational
real number).
Figure 3.8: Diagrams illustrating the quasi-periodic mode of ferroresonance
Figure 3.9 shows the explanation for the chaotic mode. The
corresponding spectrum continuous, i.e. it not cancelled for any frequency. The
stroboscopic image made up of completely separate points occupying an area in
plane v, i known as the strange attractor.
Audible noise occurs when the steel core driven into saturation.
Therefore
the core goes into a high flux density, it make due to the magnetostriction of the
steel and to the actual movement of the core laminations. Ferroresonance happen
when the noise is louder than the normal hum of transformer.
3.6.2 Overheating
Another symptom of the high magnetic field is due to stray flux heating
in parts of the transformer where magnetic flux come across into the tank wall
and other metallic parts because of the core saturated repeatedly. The effect is
bubbling of the paint on the top of the tank. Ferroresonance happen, when it has
continued extended to reason overheating of some of the larger internal
connections. This may cause damage insulation structures.
3.6.4 Flicker
METHODOLOGY
There are various types of resistors and capacitors available in the ATP
library. This can vary from linear to non-linear and capacitor braches. However, for
this work only adopted linear branch capacitor and resistor and focusing on the non-
linear inductance.
Overhead transmission lines model depends on the line length and the
highest frequency to be simulated. This simulation, the overhead transmission lines
are double circuit 400kV. The length of the overhead line circuit is approximately
37km. It represented by multi-phase models considering the distributed nature of the
line parameters due to the range of frequencies involved, phase conductors, shield
wires are explicitly modelled between towers, and only a few spans are considered.
The line parameters being clarify by line constants, using the tower structure
geometry and conductor data as input.
Transformer model can be classified into three types. The first type is matrix
representation (BCTRAN model). The second type is saturable transformer
component model to multi-phase transformer. Both type of model can be
implemented in the ATP-EMTP. The third type is topology-based models [7].
Matrix
representation
(BCTRAN
model)
will
be
used
to
represent
single
and
three‐phase,
two
and
three
winding
transformers
by
considering
the
number
of
phases,
the
number
of
windings,
the
type
of
core
(
single
phase
core,
triplex
and
three
phase
shell
type)
and
test
frequency.
The
parameter
of
open
circuit
and
short
circuit
data
can
be
obtained
from
factory
test
data.
In
addition,
clarify
the
ratings
of
the
line‐voltage,
rated
power
and
type
of
coupling.
Besides
that,
identify
the
winding
connection
such
as
an
autotransformer,
wye
and
delta.
If
the
connection
is
an
auto
transformer
or
wye,
the
rated
voltage
is
automatically
divided
by
√3
to
get
the
winding
voltage.
If
autotransformer
selected
for
primary
and
secondary
winding
(HV‐LV)
the
impedance
are
recalculate
as
shown
below
[8]:
Z*H‐L
=
ZL‐H
(
VH/
VH
–
VL)2
(4)
Z*L‐T
=
ZL‐T
(5)
Z*H‐T
=
ZL‐H
.VH.
VL
+
ZH‐T
.VH
+
ZL‐T
.VL
(6)
(
VH
–
VL)2
(
VH
–
VL)
(
VH
–
VL)
where [R] and j ù [L] are the real and the imaginary part of the branch impedance
matrix. In transient calculations can represent as;
[di/ dt] = [L]-1 [v] – [L] -1 [R] [i] (8)
For simulation of saturable cores, excitation may be omit from the matrix
description and attached externally at the model terminals in the form of non-linear
elements.
Eddy current effects, occur at the same time in a loaded transformer that
result in a non-uniform distribution of current in the conductors, and manifest
themselves as an increase in the effective resistance and winding losses with respect
to those for direct current eddy current effects in transformer windings can be model
by Foster equivalent circuits. These circuits must be of infinite order exactly to
reproduce the impedance at all frequencies. However, a computationally efficient
circuit can be deriving by fitting only at certain pre-established frequencies. A series
model of order equal or less than 2 is adequate for low-frequency transients. A
change in the magnetic field induces also eddy currents in the iron.
Because of this, the flux density will be lower than that given by the normal
magnetization curve. As frequency changes, flux distribution in the iron core
lamination also changes. For high frequencies, the flux is confine to a thin layer
close to the lamination surface, whose thickness de-creases as the frequency
increases. This indicates that inductances representing iron path magnetization and
resistances representing eddy current losses are frequency dependent. Efficient
models intended for simulation of frequency dependent magnetizing inductances
have been derive by synthesizing Cauer equivalent circuits to match the equivalent
impedance of either a single lamination or a coil wound around a laminated iron
core limb. Inductive components of these models represent the magnetizing
reactances and have to make non-linear to account for the hysteresis and saturation
effects.
5.1 Introduction
In order for the simulation work to be successfully carried out, the following
procedures were adopted:
(ii) Develop an equivalent circuit to represent the actual power transformer and
its interconnections based on actual station.
The first part of the process was to gather information on the system. This
involved finding circuit diagrams, information on transformers, capacitors etcetera
and finally the nature of the load – magnetization characteristic and type of non-
linear inductance.
The 400kV double circuits used in this simulation were taken from
Brinsworth / Thorpe Marsh (UK transmission network) design parameter [9]. The
purpose is to exhibit of ferroresonance to occurring on the system. The length of the
parallel overhead line circuit is approximately 37 km and the feeder has a 1000MVA
400/275/13kV power transformer. Figure 5.1 shows single diagram of the
Brinsworth /Thorpe Marsh circuit arrangement.
X303 X103 TR1 T10
3x 3 0x
X420
x
x
Overhead line TR2 Brinsworth
Thorpe Marsh 37km 275kV
400kV
Circuit
Disconnector Breaker
Cable
170 m
Circuit
Breaker
♦ Rin: Inner radius of one conductor. Only available with skin effect.
Figure 5.3 shows the ATP draw input window for the transmission
line/cable. A cable length of 37 km was used. Figure 5.4 shows the line
configuration.
(a) :
Selection
of
system
type
(Line/Cable)
(b) Specification
of
conductor
data
Figure
5.3:
Line/Cable
dialog
box.
(a)
Selection
of
type
(Line/Cable),
standard
data
(grounding
and
frequency)
and
model
data
(type
of
model
and
frequency)
(b)
Specification
of
conductor
data.
Figure 5.4: Line configuration
Rating 1000MVA
Type 400/275/13 kV (auto)
Core Construction Five Limb Core
Vector Yy0
Bolt main No
Bolt Yoke No
Vector% Ratio 4 & 5/Y/M 60/60/100
The transformer has been model in the BCTRAN component of ATP Draw
which use an admittance matrix representation of the form [9];
[ I ] = [ Y ]* [ V ] (9)
The elements of the matrix are deriving from open circuit and short circuit test that
made in the factory. Table 5.2 shows the data of short test factory.
The data used in this simulation model include impedances and losses are
rate at 1000 MVA (400/275/13kV) [9]. Figure 5.5 shows the ATPDraw input
window consist of the number of phases, the number of windings, the type of core
and test frequency. The BCTRAN data is based on the test report of a 1000 MVA
(400/275/13 kV) transformer shown in Table 5.1 and Table 5.2. It used
autotransformer-winding connection for the primary and secondary winding (HV-
LV). Therefore, the impedance calculated as shown in equation (4), (5), and (6).
(a) Open Circuit Data
Figure 5.5: BCTRAN dialog box of data according to Table 5.1 and Table 5.2 (a)
Open Circuit Data (b) Short Circuit Data.
Figure 5.6 shows the magnetic core characteristics curve that been
represented by peak flux linkages versus peak currents characteristic by using an
internal SATURA – routine. Subroutine SATURATION has designed to do the
conversion with some simplifying assumptions.
Figure 5.6: The saturation curve for nonlinear inductor
There are various types of resistors and capacitors available in the ATP
library. This can vary from linear to non-linear and capacitor braches. However, for
this work only adopted linear branch capacitor and resistor and focusing on the non-
linear inductance.
Current (A)
Table 5.6 to 5.8 show the value of transformer magnetizing characteristics curve 1 to
curve 4.
Figure 5.14 to 5.17 shows the saturation curve plotted using peak flux linkages as a
function of peak currents for nonlinear inductor Type 98.
Table 5.9 shows the effect of using the saturation curve 1. The output voltage
and current waveform of the red phase shown in Figure 5.18 and 5.19 but the
ferroresonance still occurs.
Table 5.9: The effect of using the saturation curve 1
Table 5.10 show the effect of using the saturation curve 2. The output
voltage and current waveform of the Red Phase had shown in Figure 5.20 and 5.21
but the ferroresonance still occurs.
Table 5.11 shows the effect of using the saturation curve 3. The output
voltage and current waveform of the Red Phase shown in Figures 5.22 and 5.23 but
the ferroresonance still occurs.
Table 5.12 show the effect of using the saturation curve 4. Figure 5.23 and
5.25 show the voltage output is reduced after 0.25 second but the current output
shows that ferroresonance still occurs.
Table 5.12: The effect of using magnetization curve 4
From the previous studies, there are a number mitigation method(s) been
used in order to dampen the occurrences of ferroresonance in the power system. In
this simulation, mitigation techniques were analysed based on the previous studies
by adding resistor at secondary side of the power transformer.
A
Figure 5.26: Simulated power transformer circuit by adding loading resistor at secondary side (A)
5.7.1 Simulation model
Figure 5.27: The output voltage waveform by adding R=1k ohm resistance, R Phase
Figure 5.28: The output voltage waveform by adding R=1kohm resistance ,Y Phase
Figure 5.29: The output voltage waveform by adding R=1k ohm resistance, B Phase
Figure 5.30: The output current waveform by adding R=1k ohm resistance, R Phase
Figure 5.31: The output current waveform by adding R=1k Ohm resistance, Y Phase
Figure 5.32: The output current waveform by adding R=1k Ohm resistance, B Phase
Figure 5.33: The output voltage waveform by adding R= 30k Ohm resistance, R
Phase
Figure 5.34: The output voltage waveform by adding R= 30k Ohm resistance, Y
Phase
Figure 5.35: The output voltage waveform by adding R= 30k Ohm resistance, B
Phase
Figure 5.36: The output current waveform by adding R= 30k Ohm resistance, R
Phase
Figure 5.37: The output current waveform by adding R= 30k Ohm resistance, Y
Phase
Figure 5.38: The output current waveform by adding R= 30k Ohm resistance, B
Phase
6.1 Conclusion
6.2 Recommendations