See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/237125284

Comparison of Simulation Tools ATP-EMTP and

MATLAB-Simulink for Time Domain Power
System Transient Studies

Article

CITATIONS READS

7 1,325

3 authors, including:

David Raisz
RWTH Aachen University
46 PUBLICATIONS 101 CITATIONS

SEE PROFILE

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

Kopernikus-Projekt ENSURE View project

All content following this page was uploaded by David Raisz on 12 December 2014.

The user has requested enhancement of the downloaded file.

Comparison of Simulation Tools ATP-EMTP and
MATLAB-Simulink for Time Domain Power System
Transient Studies

Miklós Danyek, Péter Handl, Dávid Raisz Department of Electric Power Systems
Budapest University of Technology and Economics Egry Jozsef street 18.
Hungary 1111 Budapest, Hungary
Tel: +36 14633021, Fax: +36 14633013
danyekm | handlp | raisz @vmt.bme.hu

Abstract - Continuous development of appropriate software packages makes simulation of

power engineering problems more and more effective. However, these analysis tools differ
from each other considerably from the point of view of the applicability to a special problem.
The authors compare two widespread environments: MATLAB-SIMULINK, which can be
used to simulate a wide spectrum of dynamic systems, and ATP-EMTP, which is specific
software to simulate power system transient problems. In the first part of the paper the
components (function-blocks) that can be used to build a circuit, are listed. Then three
examples are presented which demonstrate the capabilities and underline the advantages and
drawbacks of both programs.

Keywords: ATP-EMTP, MATLAB-Simulink, time domain simulation, educational use

1 Introduction
There are numerous widespread commercial software tools used by power engineers for
electrical circuit simulation purposes. It is, however, a challenging and time consuming task
to get acquainted with all the details and specialities of such a program, that's why the
majority of the users is not inclined to keep tabs on the evolution of similar products or does
not even know them.

This paper aims to compare the capabilities of two of such software packages: ATP-
EMTP/ATPDRAW and MATLAB-SIMULINK. The first one is a universal program system
for digital simulation of transient phenomena of electromagnetic as well as electromechanical
nature, whereas SIMULINK was created to model general dynamic systems using MATLAB
simulation engine.

2 Comparison of the elements of the two programs

Both programs are block-oriented simulators with a graphical user interface. In both cases
there are two types of connections between the blocks: electrical and signal wires. Former
ones link blocks like transformers, lines etc. while the latter ones connect the signal
processing units. In ATP-EMTP these blocks can be TACS or MODELS objects.
The following tables summarise the most important built-in elements (blocks) of the two
program packages. Only objects supported by ATPDraw have been considered in the ATP-
EMTP column of the next tables.
 2.1. Sources
ATP-EMTP MATLAB
DC source, current or voltage, DC voltage source,
AC source, current or voltage, AC voltage or current source,
Ungrounded AC or DC voltage source, External controlled voltage or current source
AC source, 3 phase, current or voltage, (controlled by an arbitrary signal),
Ramp source, current or voltage, 3-phase programmable control source
Two-slope ramp source, current or (time variation of amplitude, phase and
voltage, frequency by step, ramp or modulation, 2
Double exponential source, harmonics in addition).
Heidler, Standler, Cigré type source,
current or voltage,
TACS controlled source, current or
voltage.
Sources driven by arbitrary signals can be realised in both programs. Difference is in the pre-
defined, easy-to-parameterise blocks. In ATP pre-defined surge generators are available while
in MATLAB harmonics and symmetrical components can be easily generated.

2.2. Switches
ATP-EMTP MATLAB
Single phase time controlled, Single and three-phase logical controlled
Three-phase time controlled, (opens at next current zero-crossing)
Voltage controlled, ideal switch (parallel to an RC snubber
TACS (external) controlled, circuit)
Statistic (random, based on predefined
distribution functions),
Systematic (periodic).
ATP-EMTP has some specialised switches and an external controlled one. On the other hand
MATLAB has two, which can be controlled by logical functions. To model current chopping
ATP-EMTP needs only the current threshold while in MATLAB the snubber parameters have
to be adjusted.

2.3. Machines
ATP-EMTP MATLAB
Synchronous, 3 phase, Synchronous, 3 phase (Fundamental or
Synchronous with TACS control, 3 standard parameters, former ones in SI or
phase, pu),
Synchronous, set initialisation under Simplified synchronous,
ATP, 3 phase Permanent magnet synchronous,
Induction (Asynchronous), set Synchronous machine voltage regulator
initialisation under ATP, 3 phase, and exciter,
Induction (Asynchronous), set Asynchronous,
initialisation under ATP, 1 phase, DC,
DC, set initialisation under ATP. Steam turbine and governor,
Hydraulic turbine and governor.
Both software include the models of the basic electrical rotating machines with the possibility
of the correct parameter setting. MATLAB does not contain 1 phase asynchronous machine,
but it also includes steam/hydraulic turbine models and permanent magnet synchronous
machines. ATP-EMTP comprises a synchronous 3-phase machine, where the field voltage
and the mechanical power can be controlled externally (by TACS).

2.4. Lines, cables

ATP-EMTP MATLAB
Lumped Lumped
RLC equivalent 1, 2, 3 phase, Π Section line, parameters: Number of Π
RL coupled non-symmetric 2, 3, 2x3 sections and R, L, C values.
phase, Distributed (based on Bergeron's method)
RL coupled, symmetric 3, 2x3 phase. parameters given by N*N matrices,
Distributed parameters given by sequential
Transposed 1, 2, 3, 6, 2x3, 9 phase, components.
Untransposed 2, 3 phase.
LCC line/cable
defined by the geometrical and material
data of the line/cable 1…9 phase.
Bergeron, Π, J-Marti, Noda and Semlyen
type of transmission line models.
ATP-EMTP is more suitable for real and idealised simulation of the transmission lines. The
most useful type is the LCC model where the geometrical and material data of the line/cable
has to be given only. Skin effect, bundling and transposition can automatically be taken into
consideration. No such line/cable model exists in Simulink.

2.5. Transformers
ATP-EMTP MATLAB
Ideal, 1 phase (only the turn ratio can be Linear, 1 phase (2 or 3 windings),
given), Linear, 3 phase (YgY, YgD1, YgD11,
Ideal, 3 phase (only the turn ratio can be D1Yg, D11Yg, YgYD, YgYgD),
given), Saturable, 1 phase (2 or 3 windings),
Saturable, 1 phase (2 windings), Saturable, 3 phase (YgYgD),
Saturable, 3 phase (2 or 3 windings, the 12 terminal linear 3 phase transformer
winding connection and phase shift can (the connections can be set manually)
be chosen), Three phase transformers are assembled from
Saturable, 3 phase, 3-leg core type (Y/Y single-phase ones.
only) with high homopolar reluctance
BCTRAN
The transformers are defined by the parameters of the magnetising branch, the voltage, the
resistance and the inductance of the windings. The non-linear characteristic of the
magnetising branch of a saturable transformer can be given with point-pairs manually or read
from a file. MATLAB can not handle the flux linkage between legs.

2.6. Linear Elements

ATP-EMTP MATLAB
Resistor, Series RLC branch, 1 phase,
Capacitor with damping resistor, Series RLC branch, 3 phase,
Inductor with damping resistor, Parallel RLC branch, 1 phase,
 RLC in series, 1 phase, Parallel RLC branch, 3 phase.
RLC in series, 3 phase, Parameters can be defined either as R,L,C or
RLC branch 3 phase Y-connected, as P and Q.
RLC branch 3 phase ∆-connected, 3 phase mutual inductance.
Capacitor with initial voltage,
Inductor with initial current.
In this group of simulation elements the two programs are nearly equal. ATP-EMTP has two
additional elements: the capacitor/inductor with initial voltage/current. MATLAB can also
handle initial conditions, but these can be defined a bit more circuitously.

2.7. Non-linear Elements

ATP-EMTP MATLAB
Resistor, time-dependent, Metal-oxide surge arrester.
Resistor, current-dependent,
Resistor, TACS (external) controlled,
Inductor, current dependent,
Inductor, current dependent with initial
flux,
Hysteresis inductor,
Hysteresis inductor with initial flux,
Metal-oxide arrester (exponential current-
dependent resistor) 1 and 3 phase.
The dominance of ATP-EMTP is obvious due to the variety of the non-linear elements.

2.8. Power electronics

ATP-EMTP MATLAB
Diode, Diode,
Thyristor (Valve)controlled by TACS, Thyristor,
Triac. IGBT,
GTO,
Mosfet
1, 2 or 3-arm bridge of any of the former.
Power electronics elements in ATP-EMTP are ideal and only the basic parameters can be set
(ignition voltage, minimal conducting current, de-ionisation time). MATLAB includes non-
ideal ones, the parameters of that can be given more detailed and a 1, 2 or 3 arm bridge can be
defined easily.

2.9. Power electronics control blocks

ATP-EMTP MATLAB
Need to be constructed by the user using Timer (generates a control signal at
TACS or MODELS. specified transition times),
Synchronised 6 or 12-pulse generator (to
fire the 6 or 12 electronics switches of a 6
or 12 pulse converter).
 2.10. Measurement elements
ATP-EMTP MATLAB
Probe TACS, Ideal voltage and current measurement,
Ideal voltage and current measurement, Impedance measurement,
Branch voltage measurement True RMS meter,
Instantaneous power and energy (for Fourier coefficients,
most elements). THD,
Active and Reactive power
3phase sequence analyser,
abc to dq0 transformation and vice versa.
MATLAB offers a much wider range of measuring devices, which makes it simpler to
process the simulated signals.

2.11. Signal processing units

ATP-EMTP MATLAB
TACS sources in the signal processing Some additional sources:
network: Step,
DC, Ramp,
AC, Sine wave,
Pulse, Random generator (gives uniformly or
Ramp. normally distributed random signal),
Transfer functions: Arbitrary repeating sequence,
Transfer function (user defined transfer Clock,
function in the “s” domain, max. 7th From file.
order) Sinks (simulation outputs):
Integral Numerical display,
Simple derivative Time scope,
Filters (first order low/high pass) XY scope,
Devices: Power spectral density scope (FFT),
Frequency sensor Spectrum analyser (transfer function
Relay-operated switch (controlled by an between two points, with some
external signal absolute value) restrictions)
Level-triggered To File,
Transport delay (give a limited delay that Auto-correlator and cross-correlator.
consists a fix and an input dependent Continuous:
component) Derivative,
Point-by point non-linearity Integrator,
Multiple open-close switch (Gives 18 Transfer function (user defined transfer
open and close sequences at the times set) function in the “s” domain),
Controlled (resetable) integrator Variable delay (apply variable time delay
Simple derivative to the input).
Input-IF component (Output is one of the Discrete:
three inputs depending on two reference Sample/Hold (Zero/First-order sample
signals) and hold function),
Signal selector (Gives one of the inputs Transfer functions (user defined transfer
or a maximum/minimum value depended function in the “z” domain),
on a selector signal value) Math, logical, statistical:
Sample and track (Output follows the wide range of functions (complex,
 sum of the input or samples it or delays it logical, trigonometric, statistical etc.).
by ∆t) Algebraic Constraint: it solves an
Instantaneous minimum/maximum equation f(x) = 0 where x is the output of
Minimum-maximum tracking (Holds the this block, and it is indirectly connected
minimum/maximum of the inputs) back to the block's input, which is f(x).
Accumulator and counter (Holds or This connection realises the function f. In
integrates the sum of the inputs each simulation step an iteration of the
controlled by external signals) output is performed so that the input
RMS value of the sum of input signals equals to zero.
Non-linear:
Fortran statements: several types of simple non-linear
General (determined by Fortran functions (Relay, Quantiser, Dead zone,
expression) Rate limiter, Saturation etc.).
Basic Math operations n dimensions look-up table (user defined
Basic Logical operators value n-tets, e.g. in 2D case a simple
function given by pairs of points; linearly
interpolated in between),
Flip-flops (J-K, S-R, D, D Latch, Clock),
Counters.
Filters (20 types):
analog,
digital,
adaptive.
The dominance of MATLAB in the signal processing area is visible. The values of the
"Sinks" of MATLAB can be seen on-line during the simulation, and the values of most
elements can be changed on-line (by "pausing" the simulation and entering another value).
ATP-EMTP can display the results only after the simulation (if we do not consider usage of
SPY, which is not very user-friendly). MATLAB offers widespread and numerous built-in
functions, ATP-EMTP has fewer. But in both cases it is possible to build user-defined
elements with additional work.

2.12. User defined

ATP-EMTP MATLAB
Components that can be developed by the Components that can be developed by the
user programming in MODELS user either in Matlab's programming
simulation language. language or C or Fortran.
The user can define individual components in both cases.

2.13. Others
ATP-EMTP MATLAB
Transposition (Makes defined Controller blocks: PID, fuzzy, neural
transposition between interconnected 3 networks, etc.
phase elements) Optimal control toolbox
User specified (Some possibilities to load DSP blockset
the user specified elements from disk) Fixed point blockset
Frequency (Harmonic source and some Data acquisition and system
frequency dependent elements ) identification toolboxes
With certain restrictions a C code can be
 generated from the models and this code
can be compiled to an executable file.
MATLAB models can communicate with
other Windows programs over DDE or
ActiveX protocol.

3 Comparison of solution methods

ATP-EMTP uses fixed time step trapezoidal rule method for solving differential equations,
while in Simulink six types of fixed-step and eight types of variable-step solvers can be used,
depending on the complexity of the problem.

4 An example for using ATP-EMTP: sophisticated transmission line

modelling
Topic: Comparison of the insulator strings voltage stress of a Hungarian and a Japanese
overhead line due to direct lightning flashes into the towers.
In order to perform this task three main steps are necessary:
• Creating an equivalent circuit of the transmission line tower including its
grounding,
• Building a simulation model of the analysed and the neighbouring towers
including the transmission line conductors,
• Simulation of the wave processes initiated by lightning flashes using ATP-EMTP
program.
The analysed towers are plotted in Fig 1.

Hungarian 400 kV tower type IRSA Japanese 500 kV tower Nishi-Gunma line
(total height: 49.9 m, span 400 m) (total height: 120.5 m, span 650 m)
Fig 1. - The analysed towers
The equivalent circuits of the towers are determined on the basis of Japanese papers. These
papers have replaced the towers by a cylinder to calculate the tower surge impedance because
the authors have achieved good arrangement between simulations and field measurements.
Table 1. presents the parameters of the equivalent circuits of the analysed towers.
 Hungarian Japanese
H=49.9 m H=120.5 m
Z [Ohm] 132.96 124.22
H1 [m] 4.50 12.50
R1[Ohm] 10.02 10.55
L1 [microH] 3.33 8.47
Z [Ohm] 132.96 124.22
H2 [m] 9.70 20.00
R2 [Ohm] 21.60 16.88
L2 [microH] 7.18 13.56
Z [Ohm] 132.96 124.22
H3 [m] 7.10 20.00
R3 [Ohm] 15.81 16.88
L3 [microH] 5.26 13.56
Z [Ohm] 132.96 124.22
H4 [m] 28.60 68.00
R4 [Ohm] 47.42 44.31
L4 [microH] 15.78 35.59
Grounding R1=50 Ohm R2=20 Ohm
L2=5 microH

Equivalent circuit of the tower Parameters of the equivalent tower circuits

Table 1. - The equivalent circuit and its parameters of the towers

Fig 2. - The parameters of the transmission line between the towers

On the basis of the travelling wave phenomenon the neighbouring towers had to be simulated
correctly because these elements have a significant effect during lightning. The transmission
lines between the towers are defined by the electrical and geometrical parameters of the
conductors (phase conductors and shield wire). The phase conductors are bundled. J-Marti
transmission line model is used during simulations in the frequency range of 5x10-2 Hz -
5x108 Hz, taking into account the skin effect. Fig 2. shows the parameters of the Hungarian
transmission line.
The lightning flash is substituted by a HEIDLER-type surge current generator and a resistor
(Norton generator). Fig 3. shows the voltage stress of the insulator strings in each phase for
one current (lightning) flash. The effect of the reflected waves can be analysed on the basis of
the voltage curves. The main reflection points are the grounding (0.34 µs = 2x0.17 µs = 2x
travelling time of a Hungarian tower [∼50m]) and the neighbouring towers (2.66 µs = 2x1.33
µs = 2x travelling time of the span between the towers [400m]).

Hungarian tower Japanese tower

I_peak=120 kA, t_peak=2.3 µsec
5 5

[MV] [MV]

4 4

3 3

2 2

1 1

0 0

-1 -1
0 2 4 6 8 [us] 10 0 2 4 6 8 [us] 10
(file Irsa_2.pl4; x-var t) v:UPPER -X0001A v:MID -X0001B v:BOTTOM-X0001C (file Japan_1.pl4; x-var t) v:UPPER -X0001A v:MID -X0001B v:BOTTOM-X0001C

Fig 3. - The voltage stress of the insulator strings in each phase

5 An example for using MATLAB-Simulink: sophisticated PWM inverter

control modelling
This model (see Fig 4.) demonstrates a 60 Hz voltage source feeding a 50 Hz, non-linear load
through an AC-DC-AC converter. The load consists of a 50 kW resistive load parallel to a
non-linear load modelled by:
a current source of 50 A fundamental, so that cos ϕ = 0.88
a current source of 30 A 5th harmonics.

Fig 4. - Overview of the model

The 600V, 60 Hz voltage obtained at secondary of the Wye/Delta transformer is first rectified
by a six pulse diode bridge, and the filtered DC voltage is applied to an IGBT inverter
generating 50 Hz. The IGBT inverter uses Pulse Width Modulation (PWM) at a 10 kHz
carrier frequency. The circuit is discretized at a sample time of 2 µs.

The load voltage is regulated at 1 pu (380 V rms) by a PI voltage regulator using abc to dq
and dq to abc transformations (Fig 5. shows the inside of the „Voltage regulator" block seen
in Fig. 4.):

Fig 5. - Inside of the „Voltage Regulator” block

The output of the voltage regulator is a vector containing the three modulating signals used by
the PMW Generator to generate the 6 IGBT pulses. These signals can be modified so that the
inverter acts as an active filter: the correction signal is obtained by measuring the current of
the non-linear load, filtering out the 50 Hz component (see block „4th order 50 Hz elliptic
filter”) and applying a gain and a phase-shift.

The simulation is started at steady-state. At 0.04 sec the voltage set-point of the voltage
regulator is raised from 0.8 to 1, and at 0.16 sec the active filter is switched on. On Fig 6, 7, 8
and 9 one can observe the effect of these interventions.

Set-point change
Filter on

Fig 6. - The DC voltage

Of course the interactions between the 60 and 50 Hz systems (both connected to the DC bus
through a 6 pulse bridge) cause additional 300 and 360 Hz components to the DC voltage due
to the imperfect smoothening by the L1 and C elements.
 Fig 7. - Voltage at the inverter terminals

This signal consists of 10 kHz peaks, which are blurred in Fig 7.

Set-point change
Filter on

Fig 8. - Voltage of the 50 Hz supply network

Set-point change Filter on

Fig 9. - Magnitude of the 5th harmonic at the 50 Hz network

For the „Fourier" block it takes one period of 20 ms to swing in to steady-state output, this
can be seen at the beginning of the simulation and at 40 ms, where the set-point change
occurs.
6 An example realised in both environments
The following study has been realised both in ATP-EMTP and in MATLAB-Simulink (see
Fig 10. and 11.). This example is intended to demonstrate the comparison of the simulated
quantities and the calculation speed of both programs.

It consists of a 60 km transmission line connected to a 120 kV three-phase network feeding a

40 MVA inductive/resistive load supplied at 10 kV over a 40 MVA YgD11 transformer. The
capacitors located at both sides of the transformer substitute the D-winding stray capacitance
(secondary side) and the bus bar capacitance to the ground (primary side).
A single-phase fault occurs at 1/3rd of the line at 0.044 sec (1 ms before the voltage peak in
phase A). The fault is cleared by the protection at the supply side at 0.1 sec and at the remote
end at 0.2 seconds.

Fig 10. - Model in ATP-EMTP

Fig 11. - Model in MATLAB-Simulink

Fig 12. shows the current flowing from the supply and remote network to the fault. The latter
component has relatively small amplitude because it is the zero sequence return current of the
Y/∆ transformer. It can be observed that the circuit breaker located at the left side opens at the
first current zero after 0.1 sec, the right side breaker opens after 0.2 sec.
 Fig 12. – Components of the fault current (currents in phase A from the left and the right)

Primary side of the transformer Secondary side of the transformer

Fig 13. - Voltages at the primary and the secondary side of the transformer (phase A)

Fig 13. presents the voltages at the primary and the secondary side of the transformer. Both
programs yield very similar results, the difference is generally less than 0.15 %.

Deviation between the simulation results (voltages of the transformer) were normalized to the
nominal phase peak voltage and plotted in Fig 14. In case of the primary side voltage there is
a difference of more than 1.5 % at 0.2 sec, i.e. when the remote side breaker opens. These
deviations mainly result from the different solvers used to solve differential equations and the
slight differences between the models of certain elements (e.g. transmission lines,
transformers).

At the primary side of the transformer At the secondary side of the transformer

Fig 14 – Deviation between the simulation results (voltages of the transformer) carried out
with ATP and MATLAB (in percent)
 On the basis of many simulations it can be stated that the ATP calculation speed is much
faster than that of the MATLAB. In the case of the last example ATP simulation time was
15 s and MATLAB simulation took 77 seconds.

7 Conclusions
In the paper two widely used program packages for electric network simulation were
compared. It has to be pointed out that both tools are capable of simulating the same class of
problems, since it is possible to build user defined elements in both cases (though this requires
experience and is time-consuming). There are however some differences between the pre-
defined components. ATP-EMTP is designed to simulate the physical processes of
transmission lines and transformers quickly and in a convenient way but MATLAB offers
more possibilities in power electronics, signal processing and control. The paper also shows
three examples that demonstrate the mainstays of the surveyed software tools.

8 References
[1] L. Prikler, H.K. Hoidalen, ATPDraw version 3.5 for Windows 9x/NT/2000/XP User’s
Manual, SEfAS TR F5680, ISBN 82-594-2344-8, Aug. 2002.
[2] MATLAB On-line Manual
[3] H.W. Dommel, ElectroMagnetic Transients Program. Reference Manual (EMTP
Theory Book), Bonneville Power Administration, Portland, 1986
[4] Can/Am EMTP Users Group, ATP Rule Book, 1998

9 Biographies

Miklós, DANYEK was born in Budapest, Hungary, on February 28, 1976. He studied at the
Budapest University of Technology and Economics. After receiving the MSc degree in 2000 he is
a postgraduate student at the Department of Electric Power Systems. The topics of his research
are: the transients of the high voltage power systems, using similitude theory to mitigate the
spread of the experimental data about secondary arc duration, multi circuit lines secondary arcing.
He has studentship to the National Power Line Company Ltd. (OVIT).

Péter, HANDL was born in Pécs, Hungary, on July 11, 1977. He studied at the Budapest
University of Technology and Economics. After receiving the MSc degree in 2000 he is a
postgraduate student at the Department of Electric Power Systems. His main field is the transients
in the high voltage power systems, specially the transients caused by lighting flashes and the
secondary arcs. He is a holder of a scholarship at the National Power Line Company in Hungary.

Dávid, RAISZ was born in Budapest, Hungary in 1977. He graduated from the Department of
Power Systems of the University of Technology and Economics Budapest, Hungary, in June 2000
after a one year research stay at Department for Electrical Power Systems of University of
Technology Graz, Austria. He is now working towards his Ph.D. degree in Electrical Engineering.

View publication stats