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

net/publication/280736919

DYNAMIC EQUIVALENCE OF A LARGE POWER SYSTEM USING POWER

SYSTEM SIMULATOR FOR ENGINEERS (PSS/E)

Article · December 2010

CITATIONS READS

4 4,425

2 authors:

Ashik Ahmed Md. Shahid Ullah

Islamic University of Technology University of Reading
50 PUBLICATIONS 266 CITATIONS 6 PUBLICATIONS 23 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Ashik Ahmed on 07 August 2015.

The user has requested enhancement of the downloaded file.

 JOURNAL OF ENGINEERING AND TECHNOLOGY , Vol. 8 No. 2 December 2010 ISSN 1684-4114

http://www.iutoic-dhaka.edu/jet

DYNAMIC EQUIVALENCE OF A LARGE POWER

SYSTEM USING POWER SYSTEM SIMULATOR FOR
ENGINEERS (PSS/E)

Ashik Ahmed* and Md. Shahid Ullah*

Received 09 November 2010; Accepted after revision 30 December 2010

ABSTRACT
In this paper the step-by-step procedure of obtaining the network
equivalent of a large power system using Power System Simulators for
Engineers (PSS/E) is presented. Coherency among the generators of the study
system is identified using the non-linear time domain simulation obtained by
PSS/E. Generators with the most identical swing are considered to be coherent.
Dynamic aggregation of the coherent group of generators is performed based
on the Zhukov’s method. The accuracy of the procedure is demonstrated by
comparing the steady state and dynamic results of the original and the
equivalent system. The comparisons clearly indicate excellent level of accuracy
achieved from this work. The step-by-step procedure of building dynamic
equivalent presented in this paper will be extremely helpful for the researchers
to understand and work with the commercial PSS/E software.

Keywords: Dynamic equivalent, Coherency identification, Dynamic

Aggregation.

* Electrical and Electronic Engineering (EEE) Department, Islamic University of

Technology (IUT), Gazipur-1704, Bangladesh. E-mail: [email protected]
1 INTRODUCTION
One of the mostly discussed issues for the power system engineers is the
stability of power system following small or large disturbances which are
frequently occurring in any power system network. Usually the stability study is
confined within some predefined area and detail modeling of the whole power
system is not required. The portion of the power system that is not of major
concern during the study period can be reduced to an equivalent one. Equivalent
model is the reduced order model of the overall system which retains the
important features of the system. Dynamic equivalence is the process of reducing
the complexity of the system model by eliminating different generators, loads,
buses and branches.
Modal equivalent methods and coherency based methods are the two major
ways to obtain dynamic equivalent of a large interconnected power system. In
the modal equivalent method [1] the external and the study areas are defined in
an arbitrary manner. There is always a possibility of significant degradation in
the simulation results obtained using these equivalents if the study and equivalent
areas are not weakly coupled. In the coherency based methods the process of
building the dynamic equivalent comprises of three major parts, namely, (i)
identification of the coherent machines, (ii) aggregation of the coherent machines
and (iii) reduction of the network.
Coherency index is used in [2] to identify the coherent group of machines.
Based on this index, the generators are grouped if the index falls within a certain
predefined value. A comparison between slow and inertial algorithm for
generator aggregation is presented in [3]. It is shown that the slow aggregation
algorithm might yield inaccurate result when a power network is reconstructed
because of the error in the approximation in one of the coefficients. Author in [4]
has proposed an aggregation method based on structure preservation of the
coefficient matrices in nonlinear time domain representation of generating units
with detailed models. An adaptive reduction technique is proposed in [5] which
automatically determined the buses to be eliminated such that maximum
advantage of network sparsity is taken.
Different software packages are also used for the purpose of dynamic
reduction of large power system. DYNRED, a dynamic reduction program, was
developed under the ERPI project RP2447 [6]. However, the generator terminal
bus aggregation in this method produces infinite admittances linking the coherent
generator terminal buses, thereby stiffening the system. Integrated application of
a designed software tool and the CEPEL software package is presented in [7].

Journal of Engineering and Technology Vol. 8, No. 2, 2010 14

Although it is shown that the work simplifies and speeds up the overall process
of building dynamic equivalent, the user need to have the knowledge of three
additional applications besides the proposed one. Packages like EMTDC/PSCAD
can be used for the stability run but they have a limitation in the maximum
allowable size. Real time simulators like RTDS is used in [8] which can counter
the various problems that occur in real power systems but it requires the
equivalent power system with proper size due to the constraints resulting from
the capacity of the hardware/software in the simulator. The NETOMAC program
of Siemens [9] developed the evaluation equivalent method which does not
require the complete structure and parameters of the external system. The
software package PSS/E [10] is found advantageous than the aforementioned
software in the following ways:
• It can handle real large networks with more than 1,25,000 buses.
• Nonlinear time domain simulation can be performed.
• Graphical user interface enables the user to have a clearer view about the
proceedings.
• It is stand-alone software, i.e., does not need the support of other
application software.
• Good level of accuracy can be achieved.
This paper mainly presents the software PSS/E which will help the fresh
students and researchers to understand and work with this software comfortably.
Specifically, the dynamic reduction feature of the PSS/E software is discussed
step-wise. To perform the dynamic reduction on the example system the
coherency based reduction is adopted. First the external and the internal or study
system is distinguished, then the external network is reduced by PSS/E software
with a step-by-step description, then coherent machines are identified from the
study system using the non-linear time domain simulation of the system and
finally, the coherent machines are aggregated. The effectiveness of the reduction
process is validated through the comparison of non-linear time domain
simulation results of the original and the reduced system. First the methodology
for the identification of external and internal areas, step-by-step network
reduction process in PSS/E and the generator coherency identification and
aggregation method is presented. Next the results and discussion of the work is
presented which shows the effectiveness of the applied method. Finally the
conclusion is drawn.

Journal of Engineering and Technology Vol. 8, No. 2, 2010 15

2 METHODOLOGY
The overall reduction procedure is performed through the following steps:
• Distinguishing the external and study areas
• Netting generation with the load
• Building network equivalent for the external system
• Identifying the coherent groups of generators in the study system
• Getting dynamic equivalent for each coherent group

2.1 Distinguishing the external and study areas

The system portion which is not of particular interest is classified as the
external area and the remaining portion is termed as the study area. Some criteria
like bus KV levels can be set to distinguish these areas.

2.2 Netting generation with the load

In PSS/E there are several bus type codes starting from 1 to 7. Bus code 1, 2
and 3 are used to designate the PQ bus, the PV bus and the slack bus
respectively. Code 4 is used to define the inactive buses. Codes 5 to 7 are special
bus types with the following meaning:
Type 5: This is used to specify the load buses during equivalence process which
are to be retained after reduction.
Type 6: This is used to specify the generator buses during equivalence process
which are to be kept after reduction.
Type 7: This is used to specify the slack bus during equivalence process.
Before performing network equivalence or if is desired to reduce the number
of generators modeled, generators can be replaced by combining their output
with the load at the buses. Generators at all type 2 and 3 buses will be replaced
with equivalent negative load with the following exceptions:
• Buses that are designated by the user at the start of the activity.
• Buses which are indicated by type codes 6 or 7 to be retained buses.
This netting process should not hamper the pre-reduction power flow results.
The activity NETG/GNET performs the job of netting generation with the loads
in PSS/E. A snapshot view of network equivalence activity of PSS/E is shown in
Fig.1.
The steps for performing this netting process are given below:
• After opening the saved case file in the spreadsheet view in PSS/E the
activity NETG/GNET can be accessed by the path: POWER FLOW Æ
EQUIVALENCE NETWORK Æ NETG/GNET

Journal of Engineering and Technology Vol. 8, No. 2, 2010 16

 • Netting of generation can be performed either inside or outside the
selected area. The option INSIDE SELECTED SUBSYSTEM or
OUTSIDE SELECTED SUBSYSTEM has to be selected by the user.
• If selective buses are to be handled for the netting process, the selection
can be made either by subsystem or by mentioning the individual bus
names. This can be done by selecting SELECTED BUS SUBSYSTEMS
or THE FOLLOWING BUSES. If the option “THE FOLLOWING
BUSES” is selected, the user needs to specify the list of buses for which
the netting process will be conducted.

Figure 1: Snapshot of network equivalence activity in PSS/E.

Journal of Engineering and Technology Vol. 8, No. 2, 2010 17

 As the generation is converted to negative loads at the respective buses, the
type codes of all the selected type 2 and 3 will be converted to 1 after the netting
process. The diagram view of PSS/E will show newly introduced loads at the
buses. Also the load real and reactive powers are converted to negative
generation of real and reactive powers respectively.

2.3 Building Network Equivalent for the External System

An electrical equivalent is constructed by performing a reduction operation
on the admittance matrix of the external system that is to be represented by the
equivalent. In PSS/E the admittance matrix equation of the external system is
partitioned according to the following form:
⎡ I1 ⎤ ⎡Y1 Y2 ⎤ ⎡V1 ⎤
⎢ I ⎥ = ⎢Y Y ⎥ ⎢V ⎥ … … (1)
⎣ 2⎦ ⎣ 3 4⎦⎣ 2⎦
where I1 and V1 are node current and voltage at the nodes to be retained and I2
and V2 are node current and voltage at the nodes to be deleted.
The equivalent system equation will involve only I1 and V1 explicitly with
the I2 and V2 variables assumed to be linearly dependent upon I1 and V1. The
equivalent is obtained by rearranging the second row of Eq.1 as,
V2 = Y4 −1 ( I 2 − Y3V1 ) … … (2)
and substituting this into the first row of (1) to give,
I1 = (Y1 − Y2Y4 −1Y3 )V1 + Y2Y4 −1 I 2 … (3)
During the network reduction process it is to be ensured that the retained
network preserves the nature of the original network, i.e., the impact of the
deleted network has to be present on the equivalent portion. Keeping this in mind
PSS/E denotes the first term of Eq.3 as a set of equivalent branches and static
shunt elements connecting the retained nodes and the second term as a set of
equivalent currents which must be impressed on the retained nodes to reproduce
the effect of load currents at the deleted nodes. The equivalent currents may be
translated into equivalent constant real and reactive power loads at the retained
buses.
To perform the operation discussed above, the BUILD ELECTRICAL
EQUIVALENT (EEQV) activity is used. To access the EEQV activity the path is
POWER FLOW Æ EQUIVALENCE NETWORK Æ EEQV. During this EEQV
activity PSS/E performs the following steps:

Journal of Engineering and Technology Vol. 8, No. 2, 2010 18

 • External system is isolated from the study system. Buses to be retained
are identified and the selected generators are switched to negative loads,
if desired.
• Electrical equivalent of the external system is formed by performing the
matrix operation discussed in Eq.1-3.
• The equivalent of the external system is attached to the study system to
form a complete network model.

The activity EEQV creates the equivalent on the following basis:

• Buses with type code 1 are eliminated.
• Buses with type code 2 and 3 are retained.
• Type code 4 buses are ignored during the process.
• Type code 5, 6 and 7 are retained with 4 subtracted from the type code.

For the case of a three winding transformer the following issues are
considered:
• If none of the buses connecting an in-service three-winding transformer
is to be deleted, the transformer is deleted.
• If any of the connecting an in-service three-winding transformer is to be
retained, the transformer is to be equivalenced.
• All out-of-service transformers are ignored during the equivalencing
process and are deleted after the reduction process.

The converter buses of the unblocked DC lines and the sending and terminal
end buses of the FACTS (Flexible AC Transmission System) devices are
automatically retained. To differentiate the equivalenced branch from the original
branches PSS/E assigns identifier ‘99’for them. Similarly, the equivalent loads
are also assigned with load identifier ‘99’.

2.4 Identifying the coherent groups of generators

Theoretically, two generators are to be considered as coherent if the angular
difference between them keeps constant within a certain tolerance over a certain
time interval when the power system is perturbed. The mathematical definition
can be given as follows:

V&i (t ) Vi (t ) j[δi (t ) −δ j ( t )]
= e
V&j (t ) V j (t )
… … (4)

Journal of Engineering and Technology Vol. 8, No. 2, 2010 19

 Vi (0) j[δi (0)−δ j (0)]
= e
V j (0)
= const.
Considering the voltage magnitude of the coherent buses to be constant, Eq.4
can be further simplified as:
δ i (t ) − δ j (t ) = δ ij (t ) = δ ij (0) = const. … (5)
where δ i (0) and δ ij (0) = δ i (0) − δ j (0) are the initial values of the variables
calculated for the reduced model.
In this paper the machines dynamic response simulated by PSS/E following a
disturbance is used to determine the coherency of the generators. The swing
curves of the generators of the study system are observed by applying a fault and
those generators with the most identical swing curves are classified as coherent.

2.5 Getting equivalent for each coherent generator group

The dynamic equivalent of a coherent group of generating units is a single
generating unit that exhibits the same speed, voltage and total mechanical and
electrical power as the group during perturbation where those units remain
coherent.
The coherent generators of each group can be aggregated to an equivalent
one. If the voltage level of the coherent generator buses are found to be the same,
applying the Zhukov’s method [11] one can obtain the KV level for the
equivalent generator bus. In the Zhukov’s model the voltage of the equivalent
bus is defined as the average voltage of the coherent generator buses, which can
be mathematically expressed as:
n

∑V k
Vt = k =1
… … (6)
n
n

∑θ k
θt = k =1
… … (7)
n
The turns-ratio of the ideal transformer is given by:

Journal of Engineering and Technology Vol. 8, No. 2, 2010 20

 V&k
a&k =
V&t … … (8)

where V&k and V&t are the voltages at buses k and t respectively. This process is
illustrated in Fig.2 where the buses k, k-1 and k-2 represent the generator buses
in a coherent group, while the bus t indicates the equivalent bus formed by these
generator buses.
a&1 :1

a&k :1
Figure 2: Aggregation of coherent generator buses.

The electrical real and reactive power output of the equivalent generator are
the sum of the electrical power of all the individual generators in the same
coherent group. These can be represented as follows:
n
Pe = ∑ Pei
*
… … (9)
i =1
n
Qe = ∑ Qei
*
… … (10)
i =1

where Pe* , Qe* are the real and reactive power outputs of the equivalent machine
and Pei , Qei are those of the individual machines of each coherent group.
For the dynamic modeling of the multi-machine system equivalence of the
dynamic parameters have to be determined. The swing equation of the rotor is:
d ωi
Mi = Pmi − Pei − Diωi , i = 1, 2,.......n
dt … … (11)

Journal of Engineering and Technology Vol. 8, No. 2, 2010 21

where ωi , Pei , Pmi , M i and Di are the angular speed, mechanical power,
electrical power, inertia and damping constants of generator i ; n denotes the
number of coherent generators in a group. From the definition of coherency it
follows that the coherent generators have most identical angular frequencies and
thus can be assumed to be equal to ω , the swing equation can then be described
as:
n
dω n n n
(∑ M i ) = ∑ Pmi − ∑ Pei − (∑ Di )ω , i = 1, 2,.......n … (12)
i =1 dt i =1 i =1 i =1

Therefore the inertia and damping constant of the coherent generator can be
expressed as the sum of the inertia and damping constants of the coherent
generators, respectively:
n
M * = ∑ Mi ... … (13)
i =1
n
D* = ∑ Di ... … (14)
i =1

The obtained constants are of different MVA bases than the system MVA
base. So, the M and H constants are then converted to the equivalent system base
value by the following relation:
MachineMVABasei
M new,i = M i* ... (15)
SystemMVABasei

MachineMVABasei
Dnew,i = Di*
SystemMVABasei … (16)

The transient and sub-transient d and q axis reactance of the equivalent

generator can be obtained by paralleling the corresponding reactance value of all
the coherent generators:
1
X' = n
… (17)
1
∑
i =1 X i
'

These can be converted to system base by the following relation:

Journal of Engineering and Technology Vol. 8, No. 2, 2010 22

 sy ste m b a se
X d sy ste m b a se = X d m a c h in e b a se
m a c h in eb a se … … … (18)

'
The equivalent machine d and q axis time constants Tdo' and Tqo are kept as
the same as those of the original machines as each machine of the coherent group
is having equal valued time constants.

3 RESULTS AND DISCUSSION

3.1 Generation netting with load
The original system consists of 4 different areas, namely:
1. Central Operating Area (COA)
2. Western Operating Area (WOA)
3. Eastern Operating area (EOA)
4. Southern operating area (SOA)
The overall system is composed of different KV level buses like 13 KV, 33
KV, 132 KV, 230 KV, 380 KV and so on. The dynamic analysis will be
performed with a fault applied on the tie-line between area 1 and 3. So, the
external system will be composed of areas 2 and 4 whereas the study system will
be composed of areas 1 and 3. The complete external system will be represented
by one equivalent load. This reduction will be performed in PSS/E. The focus of
the study revolves around the 230 and 380 KV network buses. So the remaining
network has to be reduced, too. A partial view of the overall network is
presented in Fig.3. An overview of the original system is provided in Table 1.
The netting of generation with the load is performed in the following way:
External system
Area WOA and SOA comprise the external system. From this part of the
system only bus 21900 is kept because it is this bus through which WOA and
COA (380 KV) are interconnected. All the remaining network loads are lumped
with the generation of the respective buses. This is done by PSS/E. A total of
175 generators are netted with their loads at this stage.
Study System
Area COA and EOA comprise the study system. Here first those generator
buses are identified which are not connected to either 230 or 380 KV network
through transformers. These are the buses to be netted with their respective loads.
Total 127 buses (56 from COA and 71 from EOA) of this category are found
from the external area.

Journal of Engineering and Technology Vol. 8, No. 2, 2010 23

 Figure 3: Single line diagram of area COA (380 KV network).

Table 1: Overview of the actual system

Number of Areas 4
Number of generators 466
Number of buses 2251
Number of lines 3533
Number of transformers 1508
Number of loads 1520
Number of shunt elements
55 and 275
(switched and fixed)
Network KV levels 380, 230, 132, 115, 69, 34.5
Number of tie-lines 250

Journal of Engineering and Technology Vol. 8, No. 2, 2010 24

Next, the network equivalence process is performed. Here the information on the
generator and load buses to be kept and retained after network reduction is
required by PSS/E. This information is to be provided by the user. The way for
this is as follows:
• The type codes of the slack, generator and load buses to be retained are
to be changed by the user from 3, 2 and 1 to 7, 6 and 5 respectively on
the spreadsheet view.
OR
• Go to POWER FLOWÆ EQUIVALENCE NETWORK
• Select the tab BUILD ELECTRICAL EQUIVALNET
• Select INSIDE SELECTED SUBSYSTEM
• In the SELECT option choose THE FOLLOWING BUSES and enter the
bus numbers list.
In this case 127 generator buses (31 from COA and 96 from EOA) and 135
load buses (23 from COA and 112 from EOA) are kept after reduction.
The comparison between the original and the equivalent system obtained
after the network reduction in PSS/E is presented in Table 2.

3.2 Comparison by steady state results

To show the effectiveness of the adopted technique of network equivalence
the load flow result of the reduced and the actual system are compared. Only the
line flows through the interconnections are listed below to show the correctness
of the steady state results and for the sake of brevity. Only three interconnections
are found to be present among the areas. Table 3 clearly indicates that the
interconnection carries the same amount of real and reactive power before and
after reduction. Other line flows are also monitored and found satisfactorily
matching.

Table 2: Comparison between original and the equivalent system

Buses Branches Generators Loads Transformers
Original 2251 3353 466 1540 1508
After
network 261 410 167 124 144
reduction
Final 140 267 30 94 47
Equivalent

Journal of Engineering and Technology Vol. 8, No. 2, 2010 25

Table 3: Comparative power flow result among the actual and equivalent system
Qline
From Pline Qline
To bus (actual) Pline(equivalent)
bus (actual) (Equivalent)
no. in in MW
no. in MW in MVAR
MVAR
19011 31227 -501.8 115.0 -501.8 115.0

19019 31243 -271.8 -36.8 -271.8 -36.8

19025 21900 251.8 -13.8 251.8 -13.8

3.3 Coherency Identification

Time domain simulation is performed on the system with a three phase bolted
fault applied at 0.1 sec and cleared at 0.2 sec in the interconnection between
areas COA and EOA in order to determine the coherent generator groups. The
simulation is carried out by the dynamic simulation feature of PSS/E. The result
shows the existence of thirty families of swing curves among the study system.
Therefore, thirty coherent groups of generators are defined which include the
generators in area 1 and 3. To show the similarity in the angular deviation among
the coherent machines, one group of generator from each area is selected.

Coherent machines
65
machine at bus 31751
machine at bus 31752
60
R otor angles(degree)

55

50

45

fault applied
40 at 0.1 sec

35
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time(s)

Figure 4: Coherent machines from area 3 (at bus 31751).

Journal of Engineering and Technology Vol. 8, No. 2, 2010 26

 Coherent Machines
28
machine at bus 11301
machine at bus 11302
26
machine at bus 11303
machine at bus 11304
24

R otor angles(degree)
22

20

18
Fault applied at
16 0.1 sec

14
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time(s)

Figure 5: Coherent machines from area 1 (at bus 11301).

In Fig.4 it is seen that the two machines from area 3 are swinging coherently
by keeping a certain angular distance. The machines at bus 11301-11305 have
identical dynamic data and thus have exactly the same pattern of oscillation
following a fault. So, in this case the angular difference is zero among the
coherent group of machines which is depicted in Fig.5. Similar kind of relative
swings as shown in Figs. 4-5 are observed among the remaining 28 coherent
groups.
Lastly the dynamic aggregation of the thirty generator groups is performed as
discussed previously. The final reduced system comprises of 3 areas, 287 lines
including the transformer branches, 140 buses and 30 generators.

3.4 Comparison of dynamics result

A three phase fault is applied at 0.1 sec on the system slack bus situated in
the area-1- 380 KV network and is self-cleared after 0.1 sec. Simulation is
performed on the original and equivalent system respectively for 5 sec. From
area-3 the generator connected to the slack bus and from area-1 bus #13301 is
chosen for comparative results of the actual and the reduced system.
Figs. 6-8 and 9-11 demonstrates the comparison of terminal voltage, active
power output and rotor angle for the slack bus and the other bus respectively.
The results show good agreement among the variations of the reduced system
and the original system. The terminal bus voltage variation matches the best for
the faulted slack bus in Fig.6. Deviation in the variation of active power between
the reduced and actual system of the slack bus is also minimum as seen in Fig.7.

Journal of Engineering and Technology Vol. 8, No. 2, 2010 27

 slack bus voltage comparison
1.4
original system
reduced system
1.2 Fault applied
at 0.1 sec
Bus voltage(pu) 1

0.8

0.6

0.4

0.2
fault removed
at 0.2 sec

0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time(s)

Figure 6: Voltage responses of the faulted slack bus in the original and reduced
system.
slack bus active power comparison
7
original system
reduced system
6
fault removed
5
at 0.2 sec
Active Power(MW)

1
Fault applied
at 0.1 sec
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time(s)

Figure 7: Active Power responses of the faulted slack bus in the original and
reduced system.
Slack bus rotor angle comparison
50
original system
reduced system
45

Fault applied
at 0.1 sec
Rotor angle(degree)

40

35

30

25

20
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time(s)

Figure 8: Rotor Angle responses of the slack bus in the original and reduced
system

Journal of Engineering and Technology Vol. 8, No. 2, 2010 28

 voltage comparison
1.15
reduced
actual
Fault applied
1.1 at 0.1 sec

1.05
Vt

fault removed
0.95
at 0.2 sec

0.9
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time

Figure 9: Voltage responses of bus # 13301 in the original and reduced systems.
Active power comparison
1.35
actual
Fault applied
reduced
1.3 at 0.1 sec

1.25

1.2
Pe

1.15

1.1

1.05
fault removed
at 0.2 sec
1

0.95
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time

Figure 10: Active power responses of bus # 13301 in the original and reduced
systems.
rotor angle comparison
50
original system
Fault applied reduced system
45
at 0.1 sec
Rotor angle(degree)

40

35

30

25

20
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time(s)

Figure 11: Rotor angle responses of bus # 13301 in the original and reduced
systems.

Journal of Engineering and Technology Vol. 8, No. 2, 2010 29

However, Fig.8 suggests that there exists a little magnitude difference among the
rotor angle variations of the slack bus beyond the first swing. Similar kind of
mismatch in rotor angle phase is observed for the other machine in area-1 as
observed in Fig.11. The variation in terminal voltage and active power output for
the generator of area-1 is a little more prominent than the slack bus. Although the
peaks of the active power swings are not matching as seen in Fig.10, the
oscillations clearly depicts the same phase relation and once the oscillation dies,
they converge to the same steady state value.

4 CONCLUSIONS
A step-by-step procedure of building network equivalent of the external
system in a large interconnected power system by PSS/E is presented here. The
equivalent of the external system is obtained using PSS/E at first. The coherency
identification is performed for the internal system with the nonlinear time
domain simulation performed in PSS/E. Finally the generator aggregation is
performed using the Zhukov’s method. To show the accuracy of the dynamic
equivalent system both steady state and dynamic simulations are performed. The
presented results clearly indicate good level of accuracy achieved by the work.
This work will certainly help the researchers to perform dynamic reduction of
large power systems by commercial software like PSS/E.

REFERENCES
[1] J. M. Undrill and A. E. Turner (1971) Construction of Power System
Electromechanical Equivalents by Modal Analysis. IEEE Trans. On Power
Apparatus and Systems, Vol. PAS-90, May/June 1971, pp. 2049-2059.

[2] Jin-Yi Kim, Dong-Jun Won and Seung –II Moon (2002) A Study on the
Dynamic Reduction for Large Power System. KIEE International
Transactions on PE, 12A-1, pp. 1-5, 2002.

[3] Joe H. Chow, Pierre Accari and W. W. Price (1995) Inertial and Slow
Coherency Algorithms for Power System Dynamic Model Reduction.
IEEE Trans. on power system, Vol. 10 (2), pp. 680-685, May 1995.

[4] M. L. Ourari, L. A. Dessaint and V. Q. Do (2004) Generating Units

Aggregation for Dynamic Equivalent of Large Power Systems. IEEE
Power Engineering Society General Meeting, June 10, 2004, Vol. 2, pp.
1535-1541, Denver, CO.

Journal of Engineering and Technology Vol. 8, No. 2, 2010 30

 [5] W. F. Tinney and J. M. Bright (1987) Adaptive Reductions for Power
Flow Equivalents. IEEE Trans. on power system, Vol. PWRS-2, pp. 351-
360, May 1987.
[6] P. Kundur, G.J. Rogers, D.Y. Wong, J. Ottevangers and L. Wang(1993)
Dynamic Reduction, EPRI TR-102234 Project 2447-01, Final report, April,
1993.

[7] Adriano B. Almeida, Romeu Reginatto, Rui Jovita and G. C. da Silva

(2010) A Software Tool for the Determination of Dynamic Equivalents of
Power Systems. IREP Symposium 2010- Bulk Power System Dynamics and
Control- VIII (IREP), August 1-6, 2010, Buzios, RJ, Brazil.

[8] Y. H. Kim, S. T. Cha, J. W. Lee, T. K. Kim, J. B. Choo and H. K. Nam

(2005) Construction of Largest Equivalents Systems for Power System
Simulator. European Transactions on Electrical Power, 27 October 2005;
16: 79-91.

[9] Kang Yi, Zhou Xianlin and Xie Guoen (1998) Study of Power System
Dynamic Equivalents in NETOMAC. Power Technology, Vol. 22, No. 5,
pp. 21-24, May 1998.

[10] PTI, PSS/E Application Guide (2002). Vol. I [M]. PSS/E Brochure, 2002.

[11] Jan Machowski, Janusz W. Bialek and James RI Bumby (1997) Power
System Dynamics and Stability. John Wiley & Sons Ltd., 1997.

Journal of Engineering and Technology Vol. 8, No. 2, 2010 31

View publication stats