See discussions, stats, and author profiles for this publication at: [Link]

net/publication/224312133

Identification of problems when using long high voltage AC cable in

transmission system I: Switching transient problems

Conference Paper · May 2008

DOI: 10.1109/TDC.2008.4517183 · Source: IEEE Xplore

CITATIONS READS

7 1,314

5 authors, including:

Jacob Østergaard
Technical University of Denmark
320 PUBLICATIONS   6,846 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Offshore Energy Hubs View project

Flexpower View project

All content following this page was uploaded by Jacob Østergaard on 06 March 2014.

The user has requested enhancement of the downloaded file.

 1

Identification of Problems when Using Long

High Voltage AC Cable in Transmission System
I: Switching Transient Problems
S. Rahimi, Student Member, IEEE, W. Wiechowski, M. Randrup, J. Østergaard and A. H. Nielsen

impact [1]. Due to political and environmental pressures from

Abstract--Due to political and environmental pressures from the public and government side, upgrading and building new
the public and government side, upgrading and building new transmission facilities are becoming more and more difficult
transmission facilities are becoming more and more difficult and and in some cases the expansion of overhead transmission
in some cases the expansion of overhead transmission lines are
lines are impossible. This means that underground cable
impossible. This means that underground cable technology is the
proper substitution and solution which make the transmission technology is the proper substitution and solution which make
expansion possible with minimized visual impacts on the the transmission expansion possible with minimized visual
communities. and electromagnetic field impacts on the communities.
Within European countries, Denmark was been at the Increasing demand for electrical power by spread urban
forefront of replacing the transmission lines with cables. The centers is also leading to the use of relatively longer cable
project was supplying the power to the greater Copenhagen area circuit in HV level.
in early 1990s. Recently, the Danish TSO [Link] has
initiated a research project for analyzing the HV underground
Development of solid insulations, the optimization of cable
cable in transmission system. design and pre-moulded cable accessories contribute
The objective of this paper and the companion paper is to effectively to a lower cost of using cables in recent years. But
address the most important problems expected in transmission the cost differences between cables and overheated lines are
system with relatively larger share of long HV underground still considerable in extra high voltage (EHV) networks. This
cables. The end goal will be a guideline to special solutions and cost is inevitable for special projects in urban areas or
precautions to avoid dangerous over voltage problems and also
environmentally sensitive areas, where the construction of
resonance problems in a transmission network with future
increased share of cables. Two major categories of problems are overhead lines is rather impossible [2].
switching transient and resonance problems. In each category of Within European countries, Denmark was been at the
the possible problems, first some theoretical background is forefront of replacing the transmission lines with cables. The
provided and then the problem and countermeasures are project was to supply the power to the greater Copenhagen
discussed. In this paper switching transient problems are area in early 1990s. Recently, the Danish TSO [Link]
addressed and discussed. Some preliminary guidelines for
has initiated a research project for analyzing the HV
modeling of the components will be given which are used
especially for switching transient studies. Companion paper is underground cable in transmission system. The scope of
dedicated to address the different types of resonance problems. research is firstly to model the components of the
transmission system in detail especially the 400kV cable and
Index Terms-- Electromagnetic Transients Simulation, secondly to investigate the impact of the relatively larger
Switching Transients, High Voltage, Overhead Line, share of the cable in transmission system [3].
Underground Cable.
The objective of this paper and its companion paper [26] is
to address the second part of the research. The end goal will
I. INTRODUCTION
be a guideline to special solutions and precautions to avoid

O verhead Transmission Lines (OHL) impact on the

environment due to land occupation, land value
depreciation, effects on forests and territory, electromagnetic
dangerous over voltage problems and also resonance
problems in a transmission network with future increased
share of AC cables. Different categories of possible and most
fields (EMF), radio interference, audible noise and their visual probable problems are addressed and discussed in these two
papers. In each category of the possible problems, first some
Saeed Rahimi is Ph.D student at Centre for Electric Technology (CET) at theoretical background is provided and then the problem is
Technical University of Denmark (DTU), Ø[Link], Electrovej, Bld 325, discussed. Two main identifiable problem categories are as
DK2800 Lyngby, Denmark (e-mail: sr@[Link]) following:
Arne Hejde Nielsen is Associate Professor at CET in Technical University a) Switching Transient over-voltages
of Denmark (ahn@[Link]). Jacob Østergaard is Professor and head of b) Resonance and Harmonic Resonance
CET (joe@[Link]). Martin Randrup (MRA@[Link]) and Wojciech In this paper switching transient problems are addressed
Wiechowski (WWI@[Link]) are with “[Link]” Danish TSO.

978-1-4244-1904-3/08/$25.00 ©2008 IEEE

Authorized licensed use limited to: Danmarks Tekniske Informationscenter. Downloaded on November 10, 2009 at 03:24 from IEEE Xplore. Restrictions apply.
 2

and discussed. After defining the problem, a mathematical The variable ω shows that all quantities are calculated as
representation for cable transmission system is provided and function of frequency.
analyzed. Then some general and specific modeling guideline The electromagnetic behavior of a cable system with n
and requirements are provided, mostly based on IEEE PES parallel conductor in frequency domain is described by two
Switching Transients Task Force guidelines [4]. These matrix equations [5]:
guidelines for modeling of the component will be used in next dV ( x )
steps of the research project. Resonance and Harmonic = − Z (ω ) . I ( x) (3)
dx
Resonance problem are covered in companion paper [26].
dI ( x )
= −Y (ω ) . V ( x ) (4 )
II. DESCRIPTON OF SWITCHING TRANSIENT PROBLEMS dx
Switching transients are caused by the operation of Where V and I are vectors of dimension n representing
breakers and switches in a power system. The switching voltage and current at distance x along the cable. It should be
operations itself represent three main categories: noted that some principal assumptions are made; such as
1) Energization. longitudinally homogeneous cable system, mutually parallel
2) De-energization. axes of all n conductor, etc.
3) Re-closure. As mentioned before, all R,L,G,C can be calculated a
It is important to distinguish between first and third function of ω and geometry and material properties of cable.
category which are closely related. The energization Now the problem is calculating the voltages and currents
phenomena include energization of system elements like given in (3),(4) which are implicitly function of ω . For
transmission lines or cables, transformers, reactors, capacitor harmonic problems, i.e. case of only one specific frequency,
banks, etc with no trapped charge. But in case of re-closure, ω is angular frequency in question, and V and I represent
the line may have been left with a trapped charge after the phasors. For transient problems, V and I represent the Fourier
initial breaker opening (more pronounced in cable). In such a transforms of voltages and currents at values of ω lying
case the transient overvoltages can reach to very high values along the path of Fourier integral necessary to convert the
up to 4.0 pu [4]. De-energization includes fault clearing and solution from frequency domain to time domain. It is shown
load rejection, etc. in [7],[8] that for transient problems, ω need not be real and
Mathematical representation and analysis of the cable it is even better to follow a path of integration displaced
transmission system is a complex task and digital simulation below real axis (complex value of ω ). By eliminating the
to be employed by using one of electromagnetic transient current vector from (3), or voltage from (4) two set of the
simulation programs available. Obviously the level of detail multi conductor Telegrapher ‘s equation are obtained:
required in the model varies with the needed accuracy for the d 2V ( x )
study.
= Z (ω )Y (ω ) . V ( x ) (5)
dx 2
d 2 I ( x)
III. MATHEMATICAL REPRESENTATION AND ANALYSIS OF THE = Y (ω ) Z (ω ) . I ( x ) (6)
CABLE TRANSMISSION SYSTEM dx 2
The method of transient analysis is based on the theory of which has the below set of solutions if voltage and current at
wave propagation in multi conductor system. In switching distance of x along cable:
transient over voltage problem, it is possible to ignore the V = exp(− x ZY ).V + + exp(+ x ZY ).V − (7 )
presence of the soil and treating the problem as one classical
wave propagation along a coaxial line, with enough accuracy
for most of practical purposes. [
I = Z −1 ZY exp(− x ZY ).V + − exp(+ x ZY ).V − ] (8)
In the theory, the transmission system is defined in terms + −
V and V are vectors containing 2n constants of integration
of series-impedance matrix Z and a shunt-admittance matrix
which are resulted from system boundary condition. For
Y. It is shown in [5], [6] and [9], how these matrixes can be
studying transient problems in time domain, above solutions
derived in terms of generalized angular frequency (to present
are transferred back into the time domain by means of inverse
the frequency dependent nature of parameters) and also
Fourier transform [5], [7], [8].
geometry and material properties of cable. Latter provide a
comprehensive formulation of formulation of impedances and 1 +∞
admittances of single-core coaxial and pipe-type cables. As a V (t )=
2π
∫ V (ω ) exp( jωt ) dω (9)
result: −∞
Z (ω ) = R (ω ) + jωL(ω ) (1) Two parameters are defined from equation (8) as
following; H and Yc, the propagation and characteristic
Y (ω ) = G (ω ) + jωC (ω ) (2) admittance matrix respectively [25]:
where R,L,G,C are the series resistance, series inductance,
shunt conductance and shunt capacitance per unit length of
the cable system. These quantities are nxn matrices, where n
is the number of (parallel) conductors of the cable system.

Authorized licensed use limited to: Danmarks Tekniske Informationscenter. Downloaded on November 10, 2009 at 03:24 from IEEE Xplore. Restrictions apply.
 3

H = exp(− x ZY ) (10)
based on the traveling time t and characteristic impedance Zc
of the line and are the most efficient and accurate models
compared to the lumped parameter models (pi- circuits). In
Yc = Z −1 ZY (11) fact, the lumped parameter model contains a number of
cascaded short-sections of the line in order to represent and
Inserting them in (7),(8) gives approximate the distributed nature of the physical line. This
model needs more computational time and is less accurate.
Yc .Vk − I k = 2.H T .I mr = 2.I ki (12) Depending on the application, frequency dependency can
be incorporated. When ground return mode - zero sequence-
is involved i.e during a line to ground fault, a frequency
Yc .Vm − I m = 2.H T .I kr = 2.I mi (13)
dependent distributed parameter line model is necessary to
where k and m denote for sending end and receiving end give an accurate representation for a wide range of
voltage and current, respectively. By means of inverse Fourier frequencies. But for the cases with mostly positive sequence
transform matrix, (12) ,(13) are transferred back into the time conditions are involved (i.e, three-phase energization), simple
domain: distributed constant parameters models can provide sufficient
yc * Vk − ik = [Link] * imr (14) accuracy [4].
For calculating the line parameters of distributed parameter
model, all EMTP-type programs can be utilized. They take the
yc * Vm − im = [Link] * ikr (15)
geometrical and physical information (cable dimensions,
where yc and h are the impulse responses of Yc and height underground ground, conductor and soil resistivity, etc)
as input.
H respectively, and * denotes convolution. Now, the impulse An important issue is the limitation of using the lumped
responses yc and h are needed in (14) and (15), but it is model (pi- circuits) for short cable lengths. The reason is that
more convenient first to calculate the corresponding step in such a case the line’s traveling time t is smaller than the
responses. integration step ∆t of the simulation [10],[4]. However there
are examples that such pi-section models have been
The time domain step responses of H and Yc are
successfully used for line energization studies without
calculated from the frequency domain impulse responses excessive loss of accuracy.
using inverse Fourier Transform: Especially for the cable studies constant parameter
⎡ 1 1 +∞ 1 assumption might be too limiting. For example in a disconnect
S (ω )e jωt dω (16 )
⎤
S step (t ) = F −1 ⎢ S (ω )⎥ = ∫ switch operation case study, a frequency dependent parameter
⎣ j ω ⎦
2π
−∞ j ω model must be used because the frequencies may vary in a
where S equals either H or Yc . In order to calculate H (ω ) large bandwidth and the cable parameters significantly vary
within that range. However it seems that for solid the
and Yc (ω ) , discrete version of the inverse Fourier Transform dielectric cables, the constant parameter model is often
is used which means working in a very broad band of adequate [4].
frequencies. The presence of low frequency components As a conclusion, the existing cable model in EMTP-type
results requires a small frequency step length in order to avoid programs are accurate enough for short and medium range of
aliasing and presence of high frequency components requires cable length but the possibility for longer length to be
a high upper limit in the Fourier integral. Both result in a need examined with more detail.
to the excessively large number of frequency samples and
B. Transformers
consequently computational problems. This problem has been
overcome by the development of two special versions of the For switching surge transient studies it is better to use a
inverse Fourier Transform named StepFourier and reduced order representation with less detail compared to the
SplineFourier in [25]. very detail the transformer model used and needed for
insulation studies. As general approach, a lumped parameter
IV. MODELING GUIDELINES AND REQUIREMENTS coupled-winding model with a sufficient number of R-L-C
elements is used. This model is able to represent impedance
In this section, some general and specific modeling characteristics at the terminal within the frequency range of
guideline and requirements are provided. The equipment interest. The nonlinear characteristic of the core should be
models discussed here are typically used for switching taken into account and not the frequency characteristic of the
transient problems and include discussion of the extent of the core (except for the frequencies in order of 3-5 kHz) [4].
system to be modeled, frequency ranges and simulation time- For switching surge studies, the following approaches may
steps. The focus will be on cable transmission system. be used:
A. Cables Transmission System and OHL i ) The model directly developed from the transformer
For modeling the transmission line two well known characteristic i.e, nameplate information. Standard EMTP
approaches are utilized; distributed parameter model and models use this approach [11].
lumped parameter models. Distributed parameter models are

Authorized licensed use limited to: Danmarks Tekniske Informationscenter. Downloaded on November 10, 2009 at 03:24 from IEEE Xplore. Restrictions apply.
 4

ii ) The model obtained by measured admittance matrix represented by mean closing times, standard deviations and
from frequency response of the transformer [12],[13]. the number of times that closing operations are to be
iii ) A very detailed model obtained from the transformer performed. In practice, the probability distributions for the
geometry and material characteristics and then reduced to three breaker poles are assumed equal. Then the standard
be usable in the time domain [14],[15]. deviation for the circuit breaker poles is estimated as below:
mps
Any of above models used for modeling the system σ=
x
(19)
containing the long cable needs to be validated. A list of where mps is maximum pole span and x is an arbitrary value
practices for validation is as following: left to the user. As an example, with mps equal to 5 ms and 5
i) Comparison of frequency response of the model (within chosen for x, σ=mps/5= 1 ms give a probability of 0.012%
the desired bandwidth) with the actual characteristic of that a closing operation be performed with a pole span greater
transformer. than 5 ms. Table I show the probability values calculated for
ii) Repeating above practice for all possible open and short different values of x according to the properties of normal
circuit conditions on the windings. distribution function. See also Appendix.
iii) Fundamental frequency response in the form of open
and short circuit impedances. TABLE I
iv) Verifying turns ratio or induced winding voltages at PROBABILITY OF CLOSING OPERATION BEING PERFORMED WITH POLE SPAN
fundamental frequency.
v) Comparison of measured and computed responses of x Probability (%)
terminal capacitance, if available. 1 31,8
vi) Comparison with factory tests, if available. 2 4,6
3 0.4
C. Modeling of the Switchgear, Pre-Striking 4 0.16
In switching transient studies, the switch is usually 5 0.012
modeled as an ideal conductor (zero impedance) when it is 6 0.00059
closed, and an open circuit (infinite impedance) when open.
Various options of closing time are provided by commercial Auxiliary switch has a uniform distribution and the closing
transient programs, ranging from one-shot deterministic time for it must be uniformly spread over exactly one cycle of
closings to multi-shot statistical or systematic closings [4]. the power frequency voltage. The standard deviation for
Switches are typically categorized to two main types: auxiliary switch can be calculated based on the properties of
Statistical switches and systematic switches. Statistical uniform distributions, as :
switches are the randomly closed switches whose closing 1
times are random variables. These switches are initially open, σ= (20)
2 3f
and close at appropriate random times defined by the switch
parameters, and never open again in a selected simulation. In where f is the power frequency. The mean closing time for
Aux
return, a systematic switch can be used with closing time this switch TMean can be set in any arbitrary value not less
varied between a given beginning and ending time in steps Aux
which are uniformly spaced. A typical statistical over voltage than TClose , see Fig. 1. Also for saving simulation time, it is
study is performed by running 100 or more separate possible to set TMean to zero for all three poles.
simulations of circuit breaker closings. The peak voltages for
each closing operation can be recorded and maximum
observed over voltage can be calculated.
Generally the time dependent statistical switches are
preferable for power system circuit breakers than systematic
switch. For achieving this purpose, an auxiliary (uniformly
distributed statistical switch) and one normally distributed
time dependent statistical switch for each phase is needed. Fig
1 shows this statistical switch for only one phase or pole of
the switch.
Aux
TClose = TMean + TRandom
Aux Aux
(17 )
A
TClose = TClose + TMean + TRandom
Aux A A
(18)
It should be noted that (10) is for phase A and should be
Aux
repeated for other phases as well. TRandom is uniformly
A
distributed and TRandom is normally (Gaussian) distributed as
shown in figure A1 in Appendix. A statistical switch is Fig. 1. Schematic of Statistical Switch operation

Authorized licensed use limited to: Danmarks Tekniske Informationscenter. Downloaded on November 10, 2009 at 03:24 from IEEE Xplore. Restrictions apply.
 5

It should be noted that some modern devices can control

In the case of using breakers with pre-insertion resistors, the closing angle of the poles to close at or near the voltage
statistical parameters are applied for the initial closing with zero between the contacts (applicable to capacitor bank
the resistor. It is recommended to simulate the shorting of the switching to reduce over voltages and inrush currents). It is
resistor after a fixed time, generally up to 1 cycle [4]. necessary to consider the pre-strike during the modeling of
It is worth mentioning to introduce an important breaker operation. As an example, a user-defined component
phenomenon during the operation of switches called “Breaker has been developed to model the pre-strike effect of circuit
Pre-Strike”. During energizing of a transmission lines or other breaker closing in PSCAD [18].
equipments, traditional method of representing a circuit
breaker is to assume that the contacts can close on any part of D. Modeling of other component and Loads
the cycle. Practically and in reality, there is a closing time This section is dedicated to the guidelines for modeling of
between when the contacts start to close and when they finally some important components of power system, needed for
make. Somewhere in between, an arc may strike across the switching transient studies:
contacts as they close. This is known as “pre-strike” as • Capacitors and Inductors
shown in figure 2. The vertical axis is a measure of the • Surge arresters
withstand voltage across the circuit breaker contacts in per • Sources and Network Equivalents
unit. When the breaker is open, the withstand voltage of the • Loads
circuit breaker will be rated voltage. The time varying value Capacitors and reactors are simply modeled as lumped
of voltage across the open contacts is depicted as an absolute element. For capacitor, series inductance and loss resistance
function of the alternating voltage across the contacts. When are considered in some cases of switching transient studies i.e.
the breaker contacts is close, the separation distance between back to back switching of capacitor banks, or faults study on
the contacts reduces, then the withstand voltage reduces. capacitance bus. For more detail, damping resistance of this
Usually pre-strike occurs when the voltage across the contacts series inductance at natural frequency is taken into account.
exceeds the reducing withstand voltage of the insulating For reactors, the simple lumped inductor with a core
medium between the contacts [16-18]. saturation characteristic and series resistance is sufficient for
Open
switching transient studies. Three category of capacitance
Decreasing withstand
might be modeled in case of more detail:
voltage • parallel capacitance across the reactor (for reactor
Time
contacts
opening studies, chopping of small currents)
begin • total capacitance (consists of bushing capacitance
closing |Voltage| across and equivalent winding to ground capacitance)
Pre-strike breaker • terminal to ground and terminal to terminal
contacts
1.0 capacitance ( in case of series reactors)
Metal oxide surge arresters are represented by a nonlinear
voltage versus resistance characteristic. The mathematical
model proper for low frequency and slow front transient is as
Closed
following. The V-I characteristic of a surge arrester has
0 Time
Time contacts fully several exponential segments, where each segment can be
close approximated by
q
⎛ v ⎞
Fig. 2. Pre-strike effect in closing circuit breakers with a finite closing time [18] i= p ⎜
⎜V
⎟
⎟
(21)
Note: Speed of operation of a 230 kV, SF6 circuit breaker was found to vary ⎝ ref ⎠
with temperature, SF6 pressure, hydraulic energy and dc voltage according to a where q is the exponent, p is the multiplier for that
test undertaken by the Manitoba HVDC Research Centre as follows [18]:
Variations in closing speed: 8-10% segment, and Vref is an arbitrary reference voltage which
Variations in opening speed: 5-6%
normalizes the equation and prevents numerical overflow
One supplier provided the following circuit breaker closing times:
110 kV 30 +/- 10 msec. [19]. Also for fast front surges, the model of arrester can be
220 kV 65 +/- 8 msec. according to the circuit shown in Fig 3. A0 and A1 are two
220 kV 80 +/- 5 msec.
500 kV 80 +/- 5 msec. time independent nonlinear resistors, L0 and L1 are a pair of

As a result of the pre-strike, there will be a greater linear inductors paralleled by a pair of linear resistors ( R0 and
tendency for effective closing to occur with rising or R1 ) and a capacitor C . The V-I characteristic of A1 is
maximum voltage across the contacts. For slow contact
slightly less than the 8/20 µs curve but A0 is 20% to 30%
closing, there will even be a shadow effect where it will not be
possible any effective closing to occur over a portion of the higher. A low pass filter consists of L1 and R1 sees a decaying
cycle. voltage across it. For considering the ground leads, a lumped

Authorized licensed use limited to: Danmarks Tekniske Informationscenter. Downloaded on November 10, 2009 at 03:24 from IEEE Xplore. Restrictions apply.
 6

inductance of about 1 µH per meter is added in series with Among the available load models in the literature, there is
the model [19]. For most switching transient studies, it is not a well developed model especially used for transient
recommended that wave shape dependent characteristics are studies. Main categories of loads are as following:
not required and also the surge arrester lead lengths and • Heating and lighting loads (resistive)
separation effects can also be ignored for such studies. [20] • Active component of motor loads (resistive)
and [21] also provide more samples of practiced models. • Fluorescent lighting loads (inductive)
• Reactive components of motor loads (inductive)

In general, the power system load is represented by an

equivalent circuit with parallel-connected resistive and
inductive elements with a certain power factor. Whenever
loads are lumped at a load bus, all downstream components
contribute to the model of load such as cables, transformers,
etc. It is recommended to use an impedance in series with the
parallel R-L-C load equivalent circuit for modeling of high-
frequency transient phenomena, as shown in Fig. 5(a). As
typical value, the series impedance is in the range of 10-20%
Fig. 3. MOV surge arrester model for fast front surges [19].
of the load impedance [4]. For some specific application, it
Sources are modeled as ideal sine-wave source in might be needed to use detailed representation certain load
switching transient studies. Generators are modeled as voltage components (e.g. induction motors) as shown in Fig. 5(b).
behind sub-transient Thevenin impedance. These are Actual power system with distributed loads all over the
simplified model. Whenever more accuracy is needed, a more system may need different consideration for load modeling.
detailed model of source and power network is necessary. The load impedances also may contribute in different types of
Three typical network equivalents are shown in Fig. 4. The resonances. This issue will need much further study which is
X/R ratio in first type (a) is the representative of the damping out of scope of this paper.
in the system (typical damping angle 75%-85%). Short circuit
impedance in second type (b) represents the surge impedance
of a connected line. In case of long cable connected to the
system, this equivalent may be used. The reason is that the
reflections are not of concern in the system under study and
model (b) reduces connected lines to a simple equivalent
surge impedance. Third model (c) is applicable when
Thevenin equivalent connected system and additional
transmission lines are known, though this approach may yield
an incorrect steady-state solution if the equivalent impedance
of the parallel connected lines is of comparable magnitude to
the source impedance [4]. Three mentioned model are more
accurate near fundamental frequency and some more complex
equivalents are provided in [22],[23] to properly represent the
frequency response characteristic.

Fig. 5. Equivalent circuit representation of power system loads for simulating

switching transients
(a) Typical Load Bus [4]
(b) Composite induction motor/Static Load equivalent [24]

V. CONCLUSION
Switching transients are one of the most important
problems for operation of transmission system with large
share of AC cable. The reason is laid in the difference
between characteristics of cable and overhead line which
makes the switching transient phenomena a more severe
problem. The mathematical representation of cable
transmission system is reviewed and analyzed. Some
Fig. 4. Typical source and Network Equivalents used for switching transient preliminary guidelines for modeling of the components are
studies [4]
provided mostly based on IEEE PES Switching Transients
Task Force guidelines.

Authorized licensed use limited to: Danmarks Tekniske Informationscenter. Downloaded on November 10, 2009 at 03:24 from IEEE Xplore. Restrictions apply.
 7

VI. APPENDIX [14] F De Leon and A. Semlyen," Reduced Order Model for Transformer
Transients" IEEE Trans. PWRD, Vol. 7, No. 1, pp. 376-383, January
Figure A1 shows a normal Gaussian distribution. The 1992.
probability of covering area between specific multiples of [15] [Link], [Link] and [Link],"Response of Transformer
standard deviation is depicted in different colors. Windings to System Voltage", IEEE Trans. PAS 93 Issue. 2 pp. 457-467,
March/April 1974,

[16] D.A. Woodford, L.M. Wedepohl, “Impact of Circuit Breaker Pre-Strike on

Transmission Line Energization Transients,” in Proc. IPST’97,
International Conference on Power Systems Transients, Seattle, June 22-
26, 1997, pp250-253.
[17] [Link] and [Link], “Transmission Line Energization with
Breaker Pre-Strike”, in Proc. WESCANEX'97, Conference on
Communications, Power and Computing; Winnipeg, MB, Canada; May
22-23, 1997; pp. 105-108.
[18] Dennis Woodford, Course material in “Training Course and Workshop on
PSCAD/EMTDC”, 11 – 15 June, 2007. at [Link], .
[19] J.A Martinez, D.W. Durbak, “Parameter determination for modeling
systems transients - Part V: Surge arresters “, IEEE Transactions on
Fig. A1. Normal (Gaussian distribution) and demonstrating the probability of
Power Delivery, Vol. 20, No. 3, pp. 2073 - 2078 Jul. 2005
different areas under different multiples of standard deviation.
[20] “Modeling of metal oxide surge arresters,” Working Group 3.4.11 on
Surge Arrester Modeling, IEEE Trans. Power Delivery, vol. PWRD-7, no.
VII. REFERENCES: 1, p. 302, Jan. 1992.
[21] I. Kim et al., “Study of ZnO arrester model for steep front wave,” IEEE
Trans. Power Delivery, vol. PWRD-11, no. 2, p. 834, Apr. 1996.
[1] Comparison of Overhead Lines And Underground Cables For Electricity [22] A.S. Morched and V. Brandwajn, “Transmission Network Equivalents for
Transmission”, Joint Working Group 21/22-01, Cigre 1996 paper No. : Electromagnetic Transients Studies”, IEEE Trans. PAS, Vol. 102, No. 9,
21/22-01 pp. 2984-2994, Sep. 1983.
[2] “Background Paper Undergrounding Of Electricity Lines In Europe”, [23] A.S. Morched, J.H. Ottevangers and L. Marti, “Multiport Frequency
Commission of the European Communities, Brussels, 10 December 2003, Dependent Network Equivalents for the EMTP”, IEEE Trans. PWRD,
Available [online]: [Link]/energy/electricity/publications/ Vol 8, Issue 3, pp. 1402-1412, July 1993.
doc/[Link] [24] P. Kundur, Power System stability and control, New York: McGraw Hill,
[3] “[Link] finansierer forskning i 400 kV højspændingskabler”, 1994., p 308.
[Link], Jun 2007, Available [ online]: [Link] [25] Gustavsen.B, Sletbak.J., Henriksen.T. ”Calculation of electromagnetic
da/menu/Nyheder/Nyhedsartikler/[Link]+finansierer+forskning+i+ transients in transmission cables and lines taking frequency dependent
400+kV+højspæ[Link] effects accurately into account”, IEEE Transactions on Power Delivery,
[4] “Modeling Guidelines for Switching Transients”, IEEE PES Switching Volume 10, Issue 2, pp. 1076-1084, Apr. 1995.
Transients Task Force [Link] (Draft Copy) Contributing Members: [26] S. Rahimi, et al., “Identification of Problems when Using Long High
A.M. Gole, D.W. Durbak, E.H. Camm, [Link], R.C. Degeneff, R.P. Voltage AC Cable in Transmission System II: Resonance and Harmonic
O’Leary, R. Natarajan, J.A. Martinez-Velasco, Kai-Chung Lee, A. Resonance”, Submitted to 2008 IEEE PES Transmission and Distribution
Morched, R. Shanahan, E.R. Pratico, G.C. Thomann, B. Spherling, D.A. Conference and Exposition, 21-24 April, 2008, Chicago, IL USA.
Woodford, L. Rugeles, A. Keri....etc... Available [online]:
[Link]
[5] [Link], D.J Wilcox, “Transient analysis of underground power-
transmission systems”, Proc. IEE, Vol. 120, No. 2, pp.253-260, Feb.
1973.
[6] [Link], “Application of matrix methods to the solution of
travelling-wave phenomena in polyphase systems”, Proc. IEE, Vol. 110
Issue.12, pp. 2200-2212, Dec. 1963.
[7] N. Mullineux, S.J; Day, [Link], “Developments in obtaining transient
response using fourier transforms I. Gibbs phenomena and Fourier
integrals”, International Journal of Electrical Engineering Education,
Vol.3, Issue.1pp. 501-506, Mar. 1965.
[8] N. Mullineux, S.J; Day, [Link], “Developments in obtaining transient
response using Fourier transforms- II Use of modified Fourier transforms”,
International Journal of Electrical Engineering Education, Vol.4
Issue.1, pp.31-40, Apr. 1966.
[9] A. Ametani, “A General Formulation Of Impedance And Admittance of
Cables”, IEEE Transactions on Power Apparatus and Systems, Vol. PAS-
99,No. 3 , pp.902-910 May/June 1980
[10] H. W. Dommel., Electromagnetic Transients Program Reference Manual
(EMTP Theory Book), Report Prepared for Bonneville Power
Administration, Portland, Oregon, August 1986.
[11] [Link], H.W. Dommel, and I.I. Dommel, “Matrix Representation of
Three Phase N-Winding Transformers for Steady State and transient
Studies “, IEEE Transactions on Power Apparatus and Systems, Vol:
PAS-101, Issue: 6, pp. 1369-1378, June 1982
[12] A Morched, L. Marti, and J Ottevangers," A High Frequency Transformer
Model for the EMTP" IEEE Trans. PWRD, Vol. 8, No. 3, pp 1615-1626,
July 1993.
[13] R.J. Galarza, [Link], R.C. Degeneff, "Transformer Model Reduction
Using Time and Frequency domain Sensitivity Techniques", IEEE Trans.
PWRD, Vol. 10, No. 4, pp 1052-1059. April 1995.

Authorized licensed
View publication stats use limited to: Danmarks Tekniske Informationscenter. Downloaded on November 10, 2009 at 03:24 from IEEE Xplore. Restrictions apply.