Final-9/9/2005

EMTP Reference Models for Transmission Line Relay Testing

1. Introduction

It is well known that in order to ensure a protective relay will perform as expected; it must be
tested under realistic power system conditions. This generally means that the relay must be tested
with transient data generated from an electromagnetic transient simulation program. There are
many such programs available – ATP (Alternative Transients Program), EMTP (Electromagnetic
Transients Program) and PSCADTM (EMTDC). In each of these programs, the power system that
generates the transient data can be modeled in detail. It is also possible to easily simulate various
fault scenarios and system configurations.

The purpose of this write-up is to provide a concept for a standard transmission line test model,
its parameters, the operations that must be considered, and typical cases that are studied on the
model so that a realistic testing of a protective relay can be carried out. This write-up is not
intended to be a complete guide, but rather serve as a keystone from which to start building.

2. Basic System Model

The basic model of Fig. 1 is suitable for testing most transmission line protection applications.
The model is made up of various components – lines, transformers, sources, etc. There are three
sources in the network – S1, S2 and S3. The source angle can be varied to simulate power flows.
The transmission lines consist of one pair of mutually coupled lines (between buses 1 and 2), out
of which one is a three terminal line. Intermediate nodes are provided in the line models to
enable application of faults at various locations. Breakers and switches are also included to
simulate different configurations. This model can be expanded to include series capacitors, shunt
reactors and capacitors etc.

The physical parameters of the components that make up the network are provided in Appendix
A of this document. It is possible for the user to generate their own simulation files in the
simulation program of their choice with the provided parameters. Alternatively, the user may use
the ATP/EMTP files provided in Appendix B to simulate different configurations in the
ATP/EMTP environment.

1
 Final-9/9/2005

Fig. 1: Basic System Model for Testing Transmission Line Protection.

2.1 How to Use the EMTP Model and Test Results

A standard transmission line EMTP test model is provided that allows the user to define various
system and fault parameters, and fault location. Depending on the functionality of the transients
program employed, a number of options beyond the scope of the system model may be used.
These generally include some level of breaker and fault control. The application of these
functions will depend on what the user is trying to accomplish. One must keep in mind that a
study of the type envisaged here, is not the same as a real-time digital simulation as there is no
closed-loop control. That is, after the fault is applied the breaker simulation does not respond to
the operation of the relays being tested and automatically reconfigure the power system (open the
breaker). An EMTP study is a prerecorded event. Any simulated breaker operation or system
reconfiguration is pre-programmed in the case setup. This limitation must be recognized and the
impact on the relay performance understood.

In the event that an accurate relay model is available in the EMTP being used, then closed loop
testing with the relay module might be considered.

An EMTP case generates one or more COMTRADE files of bus voltage and line current signals
for playback to the protective relays being tested through any of a number of relay test devices.
For example, consider a phase A-to-ground fault on line L1 at mZL1 where m is equal to 0.25 and
switch SW is open. At Bus #1 the bus voltages and line L1 currents flowing out of Bus #1 are
recorded. Also recorded at Bus #2 are the bus voltages and line L1 currents flowing out of Bus
#2. This provides the data to test both line terminal relays as a system for this case within the
limitations of an EMTP study mentioned above.

2
 Final-9/9/2005

2.2 EMTP Case

The EMTP case may be as simple as applying an internal fault on the protected line of the
system represented by Fig. 1 to see if the relays at each terminal trip. Generally the case is made
up of a prefault period sufficient to reset all relay logic (i.e. loss-of-potential) and a fault period
sufficient to assure relay operation. For a two terminal line, COMTRADE files are developed
with PT and CT secondary quantities at the two line terminals.

2.3 COMTRADE Files

The COMTRADE files are digital fault records generated in a standard format that can be read
by most test sets. They consist of sampled voltage and current data, and in some cases digital
status data. The file consists of a defined pre-fault period and fault period defined by EMTP
program parameters. Some relays can generate comtrade files.

2.4 Test Sets

The test set or fault playback device usually consist of three voltage and three current amplifiers,
appropriately connected to the relay to represent the relay connection to the power system. It
also consists of control, memory and communications to allow computer control. The test sets
convert the digital sampled data to real secondary quantities that the relay would see. The
COMTRADE files are loaded to the test set with a procedure specific to the test set. The case is
played into the relay in real time and the relay is monitored for operation.

2.5 Test Results

EMTP cases can be developed for a number of internal and external fault conditions to verify
relay performance for the model system developed within the study limitations. The cases from
each terminal may be played into a relay separately or into two [or three] relays simultaneously
from two [or three] test sets operating synchronously. The test results will confirm:

• Correct relay operation . . . trips, no trips, direction, timing, fault location, targeting, outputs,
etc.
• Correct pilot system operation . . . pilot tripping, coordination of relay terminals, permissive,
blocks, etc.
• Correct installation and wiring
• Correct system functioning
• Correct settings

EMTP cases may be:

• Applied in the test lab or in the field running end-to-end satellite testing,

3
 Final-9/9/2005

• Exchanged with the manufacturer or other users to facilitate the resolution of testing or
application issues,
• Saved in a database to be used for product acceptance testing

EMTP is a powerful tool and can be used to test the relay for most applications. Its limitations,
however, in representing the interaction between the physical relay and the power system
simulator must be understood.

3. Transmission Line Models

Transmission lines play a critical role in the generation of transients and the following discussion
will cover a number of different transmission line models for use in different relaying studies.

The resistance, inductance and capacitance of overhead transmission lines are evenly distributed
along the line length. Therefore, in general, they cannot be treated as lumped elements. In
addition, some of the line parameters are also functions of frequency. For steady-state studies,
such as load flow and short-circuit studies, the only parameters needed are the positive and zero
sequence parameters calculated from tables and simple handbook formulas at the power
frequency. For electromagnetic transient studies the parameters calculated from simple formulas
are not adequate, and the line parameters must be computed using auxiliary subroutines available
in different electromagnetic transients programs. Most electromagnetic transient programs
contain two major categories of transmission line models:

1. Constant parameter models;

2. Frequency-dependent parameter models.

In the constant parameter model category, electromagnetic transients programs provide a variety
of options such as:

• Positive and zero sequence lumped parameter representation.

• Pi-section representation.
• Distributed parameter (Bergeron model) transposed and untransposed line representation.

In the frequency-dependent model category, electromagnetic transients programs may provide:

• A frequency dependent line model for transposed and untransposed lines using a constant
modal transformation matrix.
• A frequency-dependent line model for transposed and untransposed lines using a frequency-
dependent modal transformation matrix.
• Phase domain frequency-dependent line model for transposed and untransposed lines (no
modal transformation)

Since different electromagnetic transients programs have different line models available, it is not
feasible to cover all available models in this report. Most of the discussion however, concerning

4
 Final-9/9/2005

line modeling and which models are better suited for which application studies, holds true for all
transient simulation programs.

3.1 Line Models for Steady State Studies

There are a large number of steady state applications where transmission lines need to be
modeled correctly and for only one particular frequency. EMTP has a number of models that
could be used for this purpose and one must know where each model is applicable.

3.1.1 Exact-pi Circuit Model

The exact-pi equivalent circuit of a single-phase transmission line is shown in Fig. 2.

Fig. 2: Exact-pi Equivalent Circuit.

The series impedance and shunt admittance of the exact-pi equivalent circuit of a single-phase
line are given below in Equations (1) and (2):

(1)

(2)

where, are the resistance, inductance, and capacitance per unit length, is the line
length, and γ is the propagation constant which is equals to

(3)

Equations (1)-(3) show that the exact-pi circuit model can represent the line accurately at one
specific frequency.

This model is a lumped parameter model and it is good for only one frequency of interest and
one particular line length. This model includes the hyperbolic corrections with no
approximations involved and is the best model for steady-state solutions and for frequency scans.
This model takes into account the skin effect and ground return corrections. It is a multiphase

5
 Final-9/9/2005

model in the phase domain with constant R, L, C, and G of the line and it is correct for any
number of circuits in the same right-of-way. This model is not adequate for transient studies.

3.1.2 Nominal-pi Circuit Model

This model is derived from the exact-pi model described by equation 1 if the frequency or line
length is low. For overhead transmission lines this is typically the case if Km at 60 Hz,
or Km at 600 Hz. This model takes into account the skin effect and ground return
corrections. It is a multiphase model in phase domain with constant R, L, C, and G of the line
and it is correct for any number of circuits in the same right-of-way. The line is automatically
represented as untransposed, and the user could model particular transposition schemes in great
detail by cascaded connection of nominal-pi circuit models.

This model has the same limitations as the exact-pi model in addition to being limited for short
lines i.e., less than 150 kilometers at 60 Hz and less than 5 kilometers at 2 kHz. It cannot
represent frequency dependence of line parameters in frequency scans, and it cannot be used for
“electrically long” lines. This model is not good for transient studies. However, this it has been
used for transient studies by connecting a number of cascaded nominal-pi models in series.

3.2 Line Models for Transient Studies

Distributed and frequency dependent parameter models are best for transient studies. They use
traveling wave solutions, which are valid over a much wider frequency range than pi-circuit
models.

3.2.1 Nominal-pi Model

This model is not a good choice for transient studies. However, it has been used for transient
studies by connecting a number of cascaded pi-nominal models, similar to what was done in the
past with transient network analyzers. When used as such, this model has a big disadvantage of
producing reflections at the cascading points.

To adequately represent the line over various frequency ranges, a large number of nominal-pi
cascaded sections should be used. As a rule of thumb, one should use one section to represent
the line up to 100 Hz, eight sections to extend the range up to 700 Hz and 15-20 sections to
extend the range up to 1-2 kHz or keep the section lengths between 5-10 kilometers (2 kilometers
for frequency up to 5 kHz).

3.2.2 Constant-Parameter Distributed Line Model

This line model assumes that the line parameters , , and are constant. The and
are distributed and the losses are lumped in three places, which is reasonable as long as

6
 Final-9/9/2005

<< Zsurge. The above conditions are met for positive sequence parameters to
approximately 1-2 kHz, but not for zero sequence parameters. It is a good model only where the
zero sequence currents are very small, or oscillate with a frequency close to the one at which the
parameters were calculated. This frequency should not be very high to meet the condition
<< Zsurge, otherwise the line must be split into smaller sections. The transformation matrix to
decouple the propagation modes is taken as real and constant and it is less valid as asymmetry
gets stronger. The speed of execution of this model is much faster than the cascaded nominal-pi
circuit models and there are no reflections within the line. Shunt losses are ignored in this
model.

3.2.3 Frequency-Dependent Distributed Line Model

This model provides an accurate representation of the distributed nature of all line parameters, as
well as their frequency dependence. Therefore, the line parameters for this model are not
constant but functions of frequency, i.e., R(ω), L(ω), C(ω) and G(ω). All parameters are
distributed including and .

Most frequency-dependent models are based on the modal theory where multi-phase line
equations are decoupled through modal transformation matrices, so that each mode can be
studied separately as a single-phase line. The transformation matrices for untransposed or
unbalanced lines are complex and frequency-dependent. It is possible however to obtain a good
accuracy by using real and constant transformation matrices. Some programs may provide the
option of using frequency-dependent transformation matrix. However, this option has a tendency
to yield unstable solutions and should be used with diligence. The latest advancement in
frequency dependent transmission line models is the frequency-dependent phase domain model.
This model does not use the modal transformation matrix and hence does not have the
inaccuracies associated with the frequency dependence of the transformation matrix. It provides
better accuracy while dealing with asymmetrical lines.

In summary, use the pi-exact model for steady-state analysis. For transient analysis use the
frequency dependent model for the lines of main interest, and the constant frequency distributed
parameter (Bergeron) model for lines of secondary interest.

Pi-circuit model is not a good choice for transient studies. However, it has been used for
transient studies by cascading a number of nominal-pi sections, similar to what was done in the
past with transient network analyzers. When used in this way, this model has a big disadvantage
of being computationally slow and requires additional network nodes.

4. Generator Model

The following two source models are the most commonly used in the protective relaying studies:

Model 1: Ideal sinusoidal sources behind sub-transient reactances or Thevenin impedances

of the system

7
 Final-9/9/2005

Model 2: Detailed synchronous machine model

The choice of a specific model in a study depends on system configuration and the objectives of
study.

Model 1: Ideal sources with sub-transient reactances model is used for representing large
generating stations. The assumption is that the system inertia is infinite and the disturbance
under study does not cause system frequency to change. The time frame of interest is small
(approximately 10 cycles) and the machine controls such as excitation system and governor have
not responded to the disturbance. The model is commonly used in transmission line primary
relaying studies.

In a large integrated system, the system can be divided into few subsystems. Each subsystem
then can be reduced to an ideal three-phase source and equivalent positive and zero sequence
Thevenin impedances. These impedances can be calculated using a steady-state 60 Hz fault
program by isolating the subsystem from the rest of the system at the common bus between them
and then applying a fault at that bus. It is common practice to select the common buses, which
are at least one line away from the line terminals where the relaying performance is being
evaluated. Again the assumptions in using this representation are the same as those in case of
ideal sources with sub-transient reactances. However the main advantage of this model is that
the computation requirements are significantly reduced because all components within a
subsystem are reduced to a simple representation using an ideal source and equivalent Thevenin
impedances. The main disadvantage is that the Thevenin impedance represents the system
equivalence at 60 Hz only. The transient response of the system using the reduced model is not
as accurate as when the complete system is represented with all lines and sources.

Model 2: The detailed model is mostly used for representing small generating stations in non-
integrated systems for applications where the system disturbance is likely to cause change in
frequency and the relays are slow in responding to that disturbance. The model requires complete
machine data including inertia, sub-transient, transient and steady-state reactances. Models of
turbine and excitation system can also be included depending upon the time frame of study and
their response time. The detail model represents complete machine behavior from sub-transient
to steady-state time frames. Generally, the excitation and governing system are ignored for line
relaying studies. The main disadvantage is the model is complex, it requires complete machine
data, is computationally inefficient and my not provide any additional accuracy in the simulation.
It is therefore not recommended for use in large integrated system.

In the test basic system shown in Fig. 1, Sources S1 and S3 are represented by ideal sources with
Thevenin impedances behind them. Detailed model representation is used for Source S2. Model
parameters of all three sources are given in Appendix A.

5. Transformer Model
The transformer is one of the most familiar and well-known pieces of power system electrical
apparatus. Despite its simple design, transformer modeling over a wide frequency range still

8
 Final-9/9/2005

presents substantial difficulties. The transformer inductances are frequency and saturation
dependent. The distributed capacitances between turns, between winding segments, and between
windings and ground, produce resonances that can affect the terminal and internal transformer
voltages.

Models of varying complexity can be implemented in emtp for power transformers using
supporting routines or built-in models. However, none of the existing models can portray the
physical layout of the transformer, or the high frequency characteristics introduced by inter
winding capacitance effects. Although there is no single model, or supporting routine, that can be
used to represent any power transformer at every frequency range, several emtp capabilities can
be used to model any transformer type at a particular frequency range. Fundamentally power
transformers can be represented in three different ways in the emtp. They are:

• The built-in ideal transformer model.

• The built-in saturable transformer model.
• Models based on mutually coupled coils, using supporting routines.

Ideal transformers ignore all leakage by assuming that all the flux is confined in the magnetic
core. In addition they neglect magnetization currents by assuming no reluctance in the magnetic
material. The saturable transformer model eliminates these two restrictions, by considering that
around each individual coil a separate magnetic leakage path exists, and that a finite magnetic
reluctance path exists as well. The models based on matrices of mutually coupled coils can
represent quite complex coil arrangements but are somewhat more difficult to use.

The above basic transformer models do not represent saturation, eddy current, and high
frequency effects with the exception of the saturable transformer model which has saturation
built-in directly into the model. Saturation, eddy currents, and high frequency phenomena can be
represented separately by increasing the complexity of the above basic models. The above
models are used in studies where the user is interested at relatively low frequencies up to 2 kHz.
At frequencies above 2 kHz the capacitances and capacitive coupling between windings become
very important. In fact, at very high frequencies the transformer behavior is dominated by its
capacitances. Complex and detailed model is needed if one is interested in the internal transient
winding voltage distribution.

Usually transformer models are derived considering the behavior of the transformer from its
terminals. However, in relaying studies, one might be interested in internal power transformer
faults. Such a method, for the simulation of internal transformer faults, using emtp capabilities
was presented in [1]. Sometimes, if explicit representation of transformers is not required, the
user needs to model the effect of the transformer presence in the power system without the need
of any details about the transformer itself. Thevenin equivalent representations in the sequence
domain are well known and can be used in these situations.

5.1 Ideal Transformer Model

The equations that describe a single-core two winding ideal transformer are:

9
 Final-9/9/2005

(4)

(5)

where are the turns of windings 1 and 2 respectively.

The assumption in modeling ideal transformers is that all the flux is confined in the magnetic
core and that there is no reluctance in the magnetic material. The ideal transformer can be
modeled in ATP by using the type 18 source, and by setting a voltage source to zero. This
component has a very simple input format. One of its main advantages is that it can be used
together with other emtp linear and nonlinear components to represent more complex power
transformers not available in ATP.

5.2 Saturable Transformer Model

The saturable transformer model uses a star-circuit representation for single-phase transformers
with up to three windings. Its extension to three-phase units is not as accurate. This model
requires as a minimum the following information:

• The voltage rating of each winding.

• The leakage impedance of each winding, and

• The transformer connectivity information.

The leakage impedances are fixed inductances and resistances and are separated into individual
elements for each winding. The representation of the magnetizing branch is optional and is
discussed later. Impedance values in p.u. from short circuit or load flow data files, must be
converted to actual units using the following equation:

Zbase = (6)

V = voltage rating of the winding

S = single-phase power rating

The saturable transformer model is based on a classical textbook equivalent circuit of a single-
phase transformer as shown in Fig. 3.

10