UPEC2010
31st Aug - 3rd Sept 2010
Comparison of Matlab PST, PSAT and DigSILENT
for Transient Stability Studies on Parallel HVAC-
HVDC Transmission Lines
A V Ubisse K A Folly
University of Cape Town University of Cape Town
avubisse@hotmail.com Komla.folly@uct.ac.za
Abstract- In this paper, three software packages that allow
HVDC to be modeled are compared; namely: DigSILENT,
Matlab PST (Power System Toolbox) and Matlab PSAT (Power
System Analysis Toolbox). Both steady state and transient
stability studies are performed using the various software
packages. The simulation results are then compared. The
similarities and differences between the results are discussed.
The paper also looks at the modeling capabilities of the various
software packages and their limitations.
Index Terms—Steady state, Transient stability, HVAC-
HVDC transmission, DigSILENT, Matlab PST, Matlab PSAT
I. INTRODUCTION
With the increase in power demand, HVDC has become
a preferred alternative to the conventional AC transmission
system to transmit bulk power over long distances. This is
due to its economical and technical advantages in long
distance power transfer. HVDC links offer suitable solutions
for interconnecting HVAC systems with different frequencies
and it can deliver more power over longer distances with
fewer losses. HVDC systems also offer high controllability
on the power transmitted [1- 5].
When interconnected to HVAC systems, HVDC systems
can improve their performance in terms of their response to
transient stability.
Since HVDC is a relatively new technology when
compared to HVAC, not all existing software packages can
model this type of system accurately and reliably. For this
reason only a limited number of packages that allow HVDC
modeling are used in this paper. The hybrid HVAC–HVDC
transmission network is investigated.
The three software packages have been validated in [4]
by comparing the HVAC power flow results to the results
given in reference [1]. Since there is no reference against
which to compare the results of the hybrid HVAC-HVDC
network, they are compared between the software packages.
The paper is organized as follows: Section II describes
the software packages used to conduct the simulations;
section III gives a description of the system model used
across all packages. Section IV presents and discusses the
power flow results, section V investigates the impact of a
transient disturbance and section VI presents the conclusions
and limitations for the softwares.
K O Awodele D T Oyedokun
University of Cape Town University of Cape Town
Kehinde.awodele@uct.ac.za davoyedokun@ieee.org
II. SOFTWARE PACKAGES
The software packages used are briefly described below:
A. DigSILENT
DigSILENT stands for Digital Simulation and Electrical
Network calculation program and it was developed by
DigSILENT Power Factory. It is a computer aided
engineering tool that is widely used for industrial, utility,
commercial and academic applications.
DigSILENT has the ability to simulate load flow, fault
analysis, harmonic analysis and stability analysis for AC, DC
and AC-DC systems.
The load flow is performed using Newton Raphson
method [5].
DigSILENT does not however allow the user to model
the components from basic component levels, but gives a
choice of built-in configurations that are already modeled.
DigSILENT only allows fifth and sixth order generators to be
modeled, where for the sixth order used in this paper δ, ω, ψd,
ψq, ψ’’d and ψ’’q are the state variables. Simplified generator
models or higher order generator models are not available in
DigSILENT. There are 60 different types of exciter models
(from which 15 are IEEE based models) in its library and 19
different types of Power System Stabilizer (PSS) models.
Loads can be modeled as static loads (constant power,
constant current, constant impedance or a combination of the
three), or as dynamic loads (induction machines, voltage
dependant loads). The transmission lines can be modeled as
lumped π equivalent or as distributed π equivalent. Capacitors
and filters can also be modeled for reactive power
compensation, filtering harmonics and improving the voltage
in the system. In DigSILENT, HVDC systems (converter
stations) can be modeled in detail; however no modifications
can be done to the system. It allows the converter stations to
be modeled using voltage, active power, reactive power,
current, Gamma and external control.
B. Matlab PST
Power System Toolbox (PST) is Matlab run software that
was developed by Joe Chow. It allows users to model
components and performs AC and AC-DC system analysis
within Matlab environment. It consists of Matlab m-files, data
files and power system application files. It provides dynamic
models of machines and controls for performing damping
controller designs, transient and small-signal stability
simulations [6]. The generators can be modelled from the
simplest model (classical model with two state variables) to
the most complex (sixth order model which uses paper δ, ω,
E’d, E’q, ψkd and ψkq are the stave variables for the machines,).
There are only four exciter models available; namely, the
simplified model which is similar to the IEEE AC4A exciter,
however it is represented only by the amplifier characterized
by a gain KA and a time constant TA, IEEE type DC1A, IEEE
type DC2A and ST3 (which are included amongst the exciters
in DigSILENT). There are two types of Power System
Stabilizer (PSS). Type 1 uses the generator speed as input and
type 2 (using the generator’s electric power as input). The
transmission lines can only be modelled as lumped π
equivalent model. Capacitors can be modelled for reactive
power compensation and voltage boosting, but it does not
allow filters to be modelled. Loads can be modelled as static
loads (constant power, constant current, constant impedance
or a combination of the three), or as dynamic loads (induction
machines, voltage dependant loads). PST models HVDC
systems and it allows the user to modify the systems, but the
user must know the software in detail to benefit from this
option. The system converter stations are modelled using
voltage, power and inverter controls.
C. Matlab PSAT
Power System Analysis Toolbox (PSAT) is a Matlab
toolbox for static and dynamic analysis and control of electric
power systems. It was developed by Federico Milano and is
an open source software.
PSAT includes power flow, continuation power flow,
optimal power flow, small-signal stability analysis and time
domain simulation tools that can be used on AC and AC-DC
systems. It uses Newton Raphson algorithm to perform power
flow analysis. All operations can be assessed by means of
graphical user interfaces (GUIs) and a Simulink-based library
provides a user friendly tool for network design [7].
Generators of order II, III, IV, V (types I where δ, ω, e’q, e’d,
e”d, type II where δ, ω, e’q, e”q, e”d and type III where δ, ω,
ψf, ψq, ψd are the state variables), VI (using δ, ω, e’d, e’q, e”q,
e”d as state variables for the machines) and VIII can be
modelled in PSAT. Three types of exciters are available in
PSAT, namely AVR type I, AVR type II and AVR type III
none of which is found in neither PST nor DigSILENT.
Power System Stabilizer type I, II, III, IV and V are available
from the PSS library. In the load library different types of
loads such as voltage dependant, frequency dependant,
constant current, constant power and constant impedance
loads, exponential and mixed loads can be found. The
transmission lines can only be modelled as lumped π
equivalent model. Capacitors can be modelled for reactive
power compensation and voltage boosting, but it does not
allow filters to be modelled. HVDC system can be modelled
in PSAT; however no modifications can be done to the
system. Only power, voltage and current control can be
implemented in the converter stations in PSAT.
III. SYSTEM MODEL
The network shown in Fig.1 is the modified two-area,
four generator system with a parallel HVDC system, taken
from [8].
Generators have been modelled using the 6th order
machines. Each machine is rated 900 MVA set to supply 700
MW (see appendix for machine data). Two machines
connected to buses 1 and 2 are in area 1 and the other two
machines connected to buses 3 and 4 are in area 2. These two
areas are connected through a weak tie-line. For the HVAC
transmission line, Machine 3 is set to be the slack bus of this
system and all other machines are modelled as PV buses. The
machines are rated at 20 kV and transmission lines are rated
to 230 kV. There are 2 loads in the system, Load 1 = 976 +
j100 MVA in area 1 and Load 2 = 1767 + j100 MVA in area
2. To boost the voltage in the system, capacitor banks, Cap 1
= 300 Mvar and Cap 2 = 450 Mvar are connected to buses 7
and 9 receptively. Area 1 and area 2 are connected by a
double set of transmission lines that are 220 km long. The
lines are modelled as π equivalent circuits.
The exciters for this paper are the IEEE AC4A type
exciter (Fig. A1 for model used in DigSILENT and PST
(shown in the Appendix) and Fig. A2 for PSAT. PSS STAB1
shown in Fig. A3 was used for DigSILENT and PST and Fig.
A4 for PSAT. The input used in all PSS is rotor speed. Fig
A2 is simplified by canceling TE, TF and KF, i.e., setting them
to zero. The difference between the models shown Fig. A3
and the model shown in Fig. A4 is that K in Fig. A3 is equal
to K*T in Fig. A4. TE is Fig 14 is not set by the user and its
value is internally calculated from the limits of the output
signal PSAT does not allow the user to do so.
Fig. 1: Two area multi machine system with a hybrid HVAC-HVDC
transmission system
The parameters of the transmission lines are given in the
Appendix. The HVDC system is a monopolar 500 kV, 230
MVA, 0.46 kA link with 12 pulse converters on both rectifier
and inverter sides. This HVDC system is modified from the
CIGRE benchmark model [9, 10]. The HVDC line is modeled
as a π equivalent circuit. Each converter has its own
converting transformer. The HVDC transmission line is 220
km long and its parameters are given in the Appendix. To
control the power transmitted across the HVDC link the
rectifier is modelled using current control. The current
control on the rectifier station is set to 0.4kA. The power flow
results for this system are presented in tables I, II and III.
Approximately 400 MW is transferred from area 1 to
area 2 to supply load 2 that is bigger than the total generator
capacity of area 2.
 IV. POWER FLOW ANALYSIS 113 Mvar in DigSILENT and 99 Mvar and 97Mvar for PST
The steady state study was performed by analyzing the and PSAT respectively
power flow of the HVAC-HVDC system. The voltage, active Each of the HVAC lines transmits about 100 MW and each
and reactive power results of the system are shown in Tables has a loss of approximately 2 MW (2 %), and transfers close
I, II and III, respectively. to 12 Mvar. The HVDC line delivers close to 198 MW at the
It can be seen from table 1 that the voltage magnitudes inverter side of the system, and these values are similar across
are 1.03 p.u. for machines 1 and 3 are 1.01 p.u. for machines all packages. The reactive power consumed by the rectifier
2 and 4. The lowest voltage for this system is 0.98 p.u. at bus stations is similar across all packages, but differ at the
7. This value is the same for all the software packages. The inverter stations with DigSILENT absorbing 113.18 Mvar
voltages on the DC side of the rectifier and inverter are while PST absorbs 99.71 Mvar and PSAT absorbs 97.14
1.01p.u. and 1.00 p.u., respectively in DigSILENT and Mvar. The reactive power supplied by the generators in area 1
Matlab PST. Matlab PSAT does not display the voltage in Matlab is slightly higher in PST when compared to
profile for the HVDC converter stations. DigSILENT and Matlab PSAT.
The power generated by all machines with the exception
of machine 3 in the hybrid system is 700 MW. This is Table II: Active power profile
because the buses 1, 2, and 4 have been set PV buses. Active Power (MW)
Element
DigSILENT PST PSAT
Machine 3 generated roughly 703 MW because it was set as
Gen 1 700 700 700
slack bus. While the total generated real power is Gen 2 700 700 700
approximately 2803 MW for all software packages as can be Gen 3 703.07 703.36 703.09
seen in table 2, the total reactive power generated is slightly Gen 4 700 700 700
different. For example the total reactive power generated in HVAC Line 1 98.41 98.25 98.59
DigSILENT is 585 Mvar, PST is 613 Mvar and PSAT is 592 HVAC Line 2 98.41 98.25 98.59
HVDC Line 198.78 198.7 198.25
Mvar. The HVAC lines deliver close to 197 MW and absorb Rectifier 199.74 199.7 199.35
close to 18 Mvar in DigSilent, 26 Mvar on PST and 24 Mvar Inverter 198.78 198.7 198.25
in PSAT on the receiving end of the transmission lines (bus Load 1 976 976 976
9). Close to 2 MW is lost on each of the HVAC lines. The Load 2 1767 1767 1767
HVDC line transfers from the rectifier side about 199.7 MW Cap 1 0 0 0
Cap 2 0 0 0
and about 198.7 MW is received at the inverter side of the
system, and 1 MW (0.5 %) is lost on the HVDC line. The
reactive power consumed by the rectifier stations is around 83 Table III: Reactive power profile
Mvar for PST and 87 Mvar for DigSILENT and PSAT. The Element
Active Power (Mvar)
reactive power consumed at the inverter stations is close to DigSILENT PST PSAT
Gen 1 158.3 168.2 157.39
Gen 2 170.32 193.81 168.08
Table I: Voltage Profile
Gen 3 137.79 136.29 135.29
Element Rated Voltage (p.u.) and angle (deg)
Gen 4 119.2 115.17 131.11
Voltage DigSILENT PST PSAT
HVAC Line 1 9.05 12.89 12.08
(kV)
HVAC Line 2 9.05 12.89 12.08
Bus 1 20 1.03 1.03 1.03
HVDC Line 0 0 0
∠6.56° ∠6.67° ∠6.49°
Rectifier 87.3 82.87 86.77
Bus 2 20 1.01 1.01 1.01
Inverter 113.18 99.71 97.74
∠-3.14º ∠-3.04º ∠-3.2º
Load 1 100 100 100
Bus 3 20 1.03 1.03 1.03
Load 2 100 100 100
∠-6.8º ∠-6.8º ∠-6.8º
Cap 1 300 300 300
Bus 4 20 1.01 1.01 1.01
Cap 2 450 450 450
∠-16.51º ∠-16.51º ∠-16.5º
Bus 5 230 1.01 1.01 1.01
∠0.12º ∠0.21º ∠0.05º
V. TRANSIENT STABILITY
Bus 6 230 0.99 0.99 0.99
∠-9.85º ∠-9.79º ∠-9.91º Transient stability of a power system refers to the ability of
Bus 7 230 0.98 0.98 0.98
a system to remain stable, i.e., maintain synchronism, when
∠-18.05º ∠-18.07º ∠-18.11º
Bus 8 230 0.99 0.99 0.99 subjected to severe disturbances such as faults and switching
∠-24.76º ∠-25.79º ∠-24.76º of lines [1, 11].
Bus 9 230 1 1 1 The transient stability of the system is evaluated by
∠-31.24º ∠-31.24º ∠-31.2º applying a three-phase fault at bus on 8 line 8-9 at t=1s. The
Bus 10 230 1 1 1
∠-23.17º ∠-23.19º ∠-23.15º fault was cleared after 50 ms by removing the line. The
Bus 11 230 1.01 1.01 1.01 impact on all machine terminal voltages, rotor angle and
∠-13.24º ∠-13.25º ∠-13.24º active power are analysed.
Rectifier 500 1.01 1.01 * The rotor angle output in PST and PSAT are the individual
∠0º ∠0º
Inverter 500 1.01 1.01 *
machine rotor angles, and the rotor angle differences are then
∠0° ∠0° calculated using the rotor angle of machine 3 as the reference.
*: No display of the results in PSAT
 Generator terminal voltage Generator rotor angle
-0.4
Machine 1 Machine 1
Machine 2 Machine 2
1.1 -0.5
Machine 3 Machine 3
Machine 4 Machine 4
-0.6
1.05

-0.7

rotor angle [rad]

voltage [pu]

-0.8
0.95
-0.9

0.9
-1

0.85
-1.1

0.8
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
time [s]
time [s]

Fig. 2: Terminal machine voltage in DigSILENT Fig. 5: Machine rotor angle in DigSILENT

Generator terminal voltage Generator rotor angle

1.15
0.6
Machine 1
Machine 1
Machine 2
1.1 Machine 3 0.5 Machine 2
Machine 4 Machine 3
1.05 0.4 Machine 4
rotor angle [rad]

0.3
voltage [pu]

0.2
0.95

0.1
0.9
0
0.85
-0.1

0.8
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 -0.2
time [s] 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
time [s]
Fig. 3: Terminal machine voltage in Matlab PST Fig. 6: Machine rotor angle in Matlab PST

Generator terminal voltage Generator rotor angle

0.6
Machine 1 Machine 1
Machine 2 Machine 2
1.1 0.5
Machine 3
Machine 3
Machine 4
Machine 4
0.4
1.05

0.3
rotor angle [rad]
voltage [pu]

0.2
0.95
0.1

0.9
0

0.85
-0.1

0.8
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 -0.2
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
time [s]
time [s]

Fig. 4: Terminal machine voltage in Matlab PSAT Fig. 7: Machine rotor angle in Matlab PSAT
 Generator active power DigSILENT automatically calculates the rotor angle
9
Machine 1 differences for all machines with respect to the rotor angle of
8
Machine 2 machine 3.
Machine 3
Machine 4 Figs. 2, 3 and 4 show the responses of the machine voltages
7 for DigSILENT, PST and PSAT, respectively. It can be seen
from Fig. 2 and Fig 3 that the system stabilizes in less than 4
active power [pu]

6 seconds. During the fault, machine 4 reaches the lowest

voltage of 0.82 p.u. and machine 1 reaches the highest
5 voltage of 1.08p.u. The voltage curves of PSAT shown in
Fig. 4 have more oscillations and take longer to stabilize,
4
approximately 6 seconds. During the fault, machine 4 reaches
a minimum voltage of 0.86 p.u. and machine 1 reaches a
3
maximum of 1.08 when the fault is cleared.
2
Figs. 5, 6 and 7 show the results for the machine rotor
0 0.5 1 1.5 2 2.5
time [s]
3 3.5 4 4.5 5
angle differences taking machine 3 as the reference rotor
angle. The rotor angles in DigSILENT settle in less than 3.5
Fig. 8: Machine active power in DigSILENT
seconds while PST (Fig. 6) and PSAT (Fig. 7) settled at
around 5 seconds. DigSILENT displays the rotor angle with
respect to machine 3 while PST and PSAT display individual
Generators active power
9 rotor angles. For this reason the rotor angle of machine 3 is at
Machine
Machine
1
2
zero in PST and PSAT. The responses of the machine electric
8 Machine 3 powers are shown in Figs. 8, 9 and 10. In DigSILENT, the
Machine 4
electric power at machine 4 reduces to 2.3 p.u. and machine 2
7
reduces to 2.8 p.u. during the fault and after it is cleared
machine 4 reaches a maximum of 8 p.u. before stabilizing at
active power [pu]

6
3 seconds.
In PST, machine 2 shows the lowest active power at 2.3
5
p.u. and machine 4 is at 2.5 p.u. during the fault. When the
4
fault is cleared, machine 4 displays the maximum at 8.4 p.u.
and the system settles at approximately 6 seconds.
3 The electric power displays more oscillations in PSAT
when compared to the other packages. The active power at
2 machine 4 dips to 4.3 p.u. and machine 2 reduces to 4.35 p.u.
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
time [s] during the fault and once machine 1 reaches a maximum of
Fig. 9: Machine active power in Matlab PST 7.7 p.u. and the system settles after 5 seconds.

VI. CONCLUSIONS
Generator active power
9 Steady state performance and the impact of a three-phase
Machine 1
Machine 2 transient disturbance were investigated in this paper. For the
8 Machine 3 steady-state studies, across all packages the voltage and
Machine 4
active power profiles are similar and there are small
7
differences in the reactive power for the machines and
converter stations. DigSILENT and PST display similar
active power [pu]

6
behaviour for the voltage and electric power for the transient
5 disturbance. The rotor angle is stable in all packages. PSAT
displays lower voltage and electric power outputs when
4 compared to the other packages. More investigations must be
performed to fully understand the softwares packages.
3

LIMITATIONS:
2
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 In all packages, transient stability studies can be
time [s]
performed; however PSAT only allows three-phase faults to
Fig. 10: Machine active power in Matlab PSAT be modelled. PST can model line to ground, line-to-line, line-
to-line to ground and three-phase faults. It can also model
loss of a line with no fault and loss of load on a bus. Neither
converter faults nor DC line faults can be modelled in both
PST and PSAT. DigSILENT and PSAT do not allow the user HVAC line parameters:
to model the components from basic component levels. R = 0.0529 Ω/km, X = 0.529 Ω/km and B = 3.371 μS/km.
Using an Sbase = 100 MVA and a Vbase = 230 kV, the
ACKNOWLEDGEMENTS equivalent parameters in per unit values are
The authors would like to express their thanks to Famutsi R = 0.0001 p.u., X = 0.001 p.u. and B = 0.00175 p.u..
Mulumba, Paul Olulope and Severus Sheetekela from the
Power group at the University of Cape Town for their help HVDC line parameters:
and technical contribution to this paper. R = 0.0281 Ω/km, X = 0.02 Ω/km, B’ = 0.44 μS/km.
Using a Sbase = 230 MVA and a Vbase = 500 kV, the equivalent
parameters in per unit values are
REFERENCES R = 0.000025 p.u., X = 0.000022 p.u. and B = 0.000021 p.u..
[1] P. Kundur, “Power System Stability and Control”, McGraw-Hill, Inc.,
1997
[2] D. A. Woodfrod, “HVDC Transmission”, Manitoba HVDC Research
Time constants TB and TC are frequently small and are
Centre, Canada, 18 March 1998.
neglected.
[3] J. Arrilaga, Y. H. Lin, N. R. Watson, “Flexible power transmission The
HVDC options”, John Wiley & Sons, ISBN 978-0-470-05688
[4] P. Breseli, W. L. Kling, R. L. Hendriks, “HVDC Connection of
Offshore Wind Farms to the Transmission System”, IEEE Transactions
on Energy Conversion, Vol 22, No. 1, March 2007.
[5] B. K. Johnson, “The ABC’s of HVDC Transmission Technologies”,
IEEE power and energy magazine, March/ April 2007. Fig. A1: IEEE AC4A type exciter (DigSILENT and PST) [14]
[6] A V Ubisse, K A Folly, K Awodele, L Azimoh, D T Oyedokun, S P
Sheetekela, “Comparison of Matlab PST, PSAT and DigSILENT for
Power Flow Studies on Parallel HVAC-HVDC Transmission lines”, The shaded block is not included in the IEEE AC4A exciter
Proceedings of the 19th Southern African Universities Power but is given in the software packages as a component of the
engineering Conference, SAUPEC 2010, University of the exciter system.
Witwatersrand, Johannesburg.
[7] DigSILENT Power Factory Version 12.0.194 Basic User’s Manual,
2001.
[8] Joe Chow, Power System Toolbox Version 2.0 Load Flow Tutorial and
Functions Manual, 2003
[9] Federico Milano, Power System Analysis Toolbox Quick Reference
Manual for PSAT version 2.1.2, June 26, 2008
[10] G. Rogers, “Power system Oscillations”, Kluwel Academic Publishers,
2000
[11] M. O. Faruque, Y. Zhang and V. Dinavahi, “Detailed Modelling of the
CIGRE HVDC Benchmark System Using PSCAD/EMTDC and
PSB/SIMULINK”, IEEE Transactions on Power Delivery, Vol. 21, No. Fig. A2: Exciter type II (PSAT)
1, January 2006
[12] Working Group 14.02, “The CIGRE benchmark model – A new
proposal with revised parameters”, December 2003.
[13] K. R. Padiyar, “Power System Dynamics Stability and Control”, John
Wiley & Sons (Asia) Pte Ltd and Interline Publishing Pvt. Ltd, 1996.
[14] “IEEE Recommended Practice for Excitation System models for Power
System Stability Studies”, IEEE Power Engineering Society, 2005
Fig. A311: Power System Stabilizer (DigSILENT and PST)

APPENDIX
Synchronous generator data
Xd = 1.8 X’d = 0.3 X’’d =0.25 Xq = 1.7 X’q = 0.55
X’’q = 0.25 Xl = 0.2 Ra = 0.0025 T’do = 8s T’’do = 0.03s Fig. A4: Power System Stabilizer type II (PSAT)
T’qo = 0.4s T’’qo = 0.05s Asat = 0.015 Bsat = 9.6 ψTI = 0.9
S1.0 = 0.039 S1.2 = 0.223 KD = 0 H = 6.5(For machines 1
and 2) H = 6.175(for machines 3 and 4)

Exciter data
KA = 200 TA = 0.05s TR = 0.01s

Power System Stabilizer data

K = 20 T = 10s
T1 = 0.05s T2 = 0.02s T3 = 3s T4 = 5.4s
The PSS washout gain (K) in PST and DigSILENT is
multiplied by the washout time constant (T) to be the same as
in PSAT.