International Journal of Engineering & Technology, 7 (4.

44) (2018) 22-27

International Journal of Engineering & Technology

Website: www.sciencepubco.com/index.php/IJET

Research paper

Analysis of Transient Stability in Synchronous Generator Using

Nyquist Method
Mohammad Noor Hidayat1, Syarifatul Izza1*, Ratna Ika Putri1

State polytechnic of malang

Electrical Engineering Department
*Corresponding author E-mail: [email protected]

Abstract

There are two factors that must be considered for the power quality analysis of an electrical power system, i.e. frequency and voltage
stability. The system that can return to the initial condition or the steady state after a disturbance occurred indicates that the system is
stable. This paper aims to detect the transient stability of three generators, which have identical specification of 2.25 MW 6.6 kV and are
operated in parallel, based on the frequency performance. In this paper, the stability analysis is determined by using the nyquist method,
where the system stability is indicated by the value of the response curve of the system. The analysis results show that the system of three
paralleled generators have the value that far from 1 + j0 and the phase margin value and gain margin tend to be minimum. These values
indicate that the system is unstable. The system slowly changes into more unstable because the oscillations are getting bigger until the
final state. So, it is necessary to change the parameters of each generator that can affect the system stability during synchronization state.

Keywords: Transient stability, synchronous generator, nyquist method, power system stability

ty and error when measuring. The previous works [4][5] investi-

1. Introduction gate one or more generators that operate separately, not in parallel
connection.
Electric power system problems include power losses, power sta- This paper describes a case study on Pertamina EP where there are
bility, losses synchronism, voltage drop, contingency etc. Power three generators with the same specification connected in parallel
stability is the ability of a system to be able to restore conditions to one bus. The transient stability is investigated by varying gov-
after an interruption occurs in the system [1]. For example, there is ernor settings and load changes using ETAP software and the
a sudden release of large loads, the release of the generator from modeling of three synchronous generators to investigate the sys-
the interconnection system which causes a large voltage drop in tem stability through the Nyquist method is generated using
the system, and a short circuit on the bus or transformer. Power MATLAB software. Frequency response methods is easier than
stability is divided into three classifications: voltage stability, the root placement method, especially for large-scale systems, and
angle stability, and long-term stability [1]. Whereas angle stability it can be made accurate with the availability of sinus generators
is further divided into small signal stability and transient stability. and expanding low frequencies which are important in a system
Transient stability is the ability of a system to maintain synchroni- according to the concept of this research.
zation when the system experiences transient interference [1].
When a synchronous machine fails or falls (falls out step), the
rotor will spin faster or decrease the speed needed to generate the
2. Power system stability
system voltage. Loss of synchronization can also occur between
one generator and another. A system that allows it to remain in its initial condition and can
The stability of the synchronous generator can be determined by return to its initial condition if there is a disturbance in the power
using Nyquist method, Routh analysis, or Lyapunov analysis [2]. system is indicated as a stable system. The analysis can use the
In this paper, the authors choose the Nyquist method. By using ideal model in detecting properties in power system instability.
Nyquist stability criterion, a plant that cannot be properly charac- The followings are two types of system stability, i.e. rotor angle
terized, can be handled. Stability conditions can be determined stability and transient stability.
from the results of the frequency response (performance in the
frequency domain) and the response in the time domain [3]. 2.1. Rotor angle stability
The previous research [4] shows that the system stability is deter-
mined by using the pole transfer on the power system stabilizer The system stability depends on the component of torque for each
through dampening the plot and the GEP method. The result, i.e. synchronous machine, include lack of torque during synchroniza-
the Power System Stability (PSS) gain cannot support the GEP tion that may cause instability due to aperiodic drift at rotor angle.
method because the method does not produce direct effects when Or, lack of damping torque as the results of oscillation instability
setting excitation. Another research [5] shows that the system The following is the explanation of the system stability consider-
stability is determined by adding impedance between source and ing the rotor angle.
load. In this case, it is necessary to pay attention to noise sensitivi-
Copyright © 2018 Authors. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted
use, distribution, and reproduction in any medium, provided the original work is properly cited.
International Journal of Engineering & Technology 23

2.1.1. Small signal (or small disturbance ) stability

Small disturbances in synchronization are caused by small varia-

tions in load and generation. The response of the system in a small
disturbance depends on some factors, include the initial operation,
the transmission system strength, and the type of generator excita-
tion controls used. For instability in a generator that is connected
radially to a large power system, it is determined by a lack of ade-
quate synchronization torque. This instability is gained through a
non-oscillatory mode, as shown Fig.1. For continuous operation of
voltage regulators, the small-disturbance stability problem can be
handled by ensuring sufficient damping of system oscillations.
Instability is normally caused by oscillations of increasing ampli- Fig. 3 Rotor angle response to a transient disturbance [1]
tude. Fig.2 illustrates the nature of generator response with auto- The behavior of a synchronous machine for a stable and unstable
matic voltage regulator [1] situation illustrated in figure 3. It shows a stable response and two
unstable response of the rotor angle. For the stable case (case 1),
the rotor angle increases to a maximum. After reaching the maxi-
mum value, the response moves into the steady state value by
decreasing the amplitude (forming an oscillation waveform) until
it reaches a steady state. In case 2, the rotor angle keeps increase
until it lost the synchronism. In case 3 the system is stable in the
first swing, but as case 2, the response becomes unstable, because
of the oscillation of the response is increasing until it lost the syn-
chronism. This unstable condition is not necessarily as a result of
the transient disturbance, but because of the instability "small-
signal" that generally occurs in the postfault steady-state condition
itself [1].

2.2. Frequency response

Fig. 1 Nature of small disturbance response with constant field voltage [1] The basis of Nyquist’s original work on the stability of feedback
amplifier is a Frequency response. Frequency response can be
calculated quite easily, if tediously, from state space equations and
a powerful tool for control system design.
Given a state space system (A, B, C, D) the frequency response
may be calculated by replacing d/dt by Iω and eliminating the
state variable x to give [1].

𝑖𝜔 𝑥 = 𝐴𝑥 + 𝐵𝑑
𝑥 = (𝑖𝜔𝐼 − 𝐴)−1 𝐵𝑑
𝑦 = 𝐶𝑥 + 𝐷𝑑
= (𝐶(𝑖𝜔 𝐼 − 𝐴)−1 𝐵 + 𝐷)𝑑 (1)

To provide the required frequency response, it is necessary to

calculate the gain margin and the phase margin. If the frequency
setting is not available, the system cannot be controlled and ob-
served. Frequency response can be observed well if using mathe-
Fig. 2. Nature of small disturbance response with excitation control [1] matical model validation to be used in the control system and
stability analysis
2.1.2. Transient stability

The ability of a system to return to its normal state when a fault

occurs is called transient stability. The system stability depends on
the initial operation and the high disturbance. In general, the sys-
tem is changed in a steady state condition after disturbance occurs
so that the system can return to normal condition. The severity of
the disturbance usually considered are short-circuit of different
types: phase-to-ground, phase-to-phase-to-ground, or three-phase.
They are usually assumed to occur on transmission lines, but oc-
casionally bus or transformer faults are also considered. The fault Fig. 4: Close loop transfer function [1]
is assumed to be cleared by the opening of appropriate breakers to
isolate the faulted element. In some cases, this process is assumed Nyquist used a method, called the principle of the Argument to
to be fast and appropriate. determine the number of zeros of a transfer function which lie in
the right-hand half plane, provided that the number of poles if the
transfer function which lie in the right-hand half plane are known.
The method is based on Cauchy’s residue theorem [1].

∮𝑐 𝑤(𝑧)𝑑𝑧 = 2𝜋 𝑖 ∑ 𝑟𝑒𝑠𝑖𝑑𝑢𝑒𝑠 𝑜𝑓 𝑤(𝑧)𝑤𝑖𝑡ℎ𝑖𝑛 𝑐 (2)

24 International Journal of Engineering & Technology

C is any closed contour in an Argand diagram and the contour is the analysis may be carried out computing only the half-part relat-
traversed in the anti-clockwise direction. Consider a transfer func- ed to the positive frequencies. The Nyquist stability criterion es-
tion having the form tablishes that Pc = P0 + N [7–8], where N is the number of clock-
wise encirclements of the point (−1,0) in the complex plane made
𝑍
∏1 𝑛(𝑠−𝑧𝑖 )𝑎𝑖 by the NP and Po is the number of unstable poles of the OLTF,
𝑇(𝑠) = 𝑃 𝑈(𝑠) (3)
∏1 𝑛 (𝑠−𝑝𝑘 )𝑏𝑘 while Pc is the number of unstable poles of the CLTF. For the NP
with positive frequencies, only the poles with positive imaginary
Where: part are considered. Assume that a transfer function G(s) has an
the zi are zeros in the right hand half of the s plane, unstable complex conjugated pair of poles o± jωo, with o > 0. In
the pk are poles in the right hand half of the s plane, and U(s) is this case, Po = 1 and Pc must be made equal to zero by designing
analytic in the whole of the right hand half of the s plane a feedback controller H(s) so the compensated OLTF encloses the
point (−1,0) of the complex plane in the counter clockwise direc-
Now, if we take the natural log of T(s) and differentiate with re- tion (N= − 1). Therefore, the design of H(s) for feedback stabiliza-
spect to s we get tion, is focused on ensuring the mapping of the compensated
OLTF encircles the point (−1,0) in the counter clockwise direction
𝑑 ln(𝑇(𝑠)) 𝑍𝑛 𝑎𝑖 𝑃𝑛 𝑏𝑘 𝑈 ′ (𝑠) [7-8].
𝑊(𝑠) = = ∑𝑖=1 − ∑𝑘=1 + (4)
𝑑𝑠 𝑠−𝑧𝑖 𝑠−𝑝𝑘 𝑈(𝑠) Otherwise, for closed-loop stability situation, the poles must have
a negative real part, which corresponds to the roots (i.e., zeros) of
The residues of W(s) are the multiplicities of the zeros and poles the characteristic polynomial 1+ kL(s). The poles of 1 + kL(s)
of T(s) in the right hand half of the s plane. which also the poles of L(s), construct the region D as a D-shaped
We next integrate W(s) around a specific contour (shown in Fig- region that containing an arbitrarily large part of the complex
ure 8) that encloses the whole of the right hand half of the s plane, right-half plane. Note that even the amount of it part was large, but
i.e it still a finite unit.
Note that s moves along the boundary of this region, 1 + kL(s)
𝑍𝑛 𝑃𝑛
∮𝑐 𝑊(𝑠)𝑑𝑠 = −2𝜋 𝑖(∑𝑖=1 − ∑𝑘=1 𝑏𝑖 ) (5) encircles the origin N = Z − P times, where
i. the number of the unstable closed-loop poles denoted as
In this case, because of the right-hand side of 3.18 is negative, the Z (zeros of 1 + kL( s) in the rhp);
contour is traversed in the clockwise direction. Because of its ii. the number of unstable open-loop poles denoted as P
analytic in the right-hand half s plane, so the contribution term of (poles of 1 + kL( s) in the rhp);
U’(s)/U(s) doesn’t count. There are no poles in the result of the
right-hand plane. Now let T(s) = rew, then Note that s moves along the boundary of this region, L(s) encircles
the −1/k point N = Z − P times, where
𝑑 ln(𝑇(𝑠)) 𝑑𝑟 𝑑𝜃 a. the number of unstable closed-loop poles denoted as Z
= + 𝑖 (6)
𝑑𝑠 𝑑𝑠 𝑑𝑠 (zeros of 1 + kL( s) in the Nyquist contour);
b. the number of unstable open-loop poles denoted as P
and (poles of 1 + kL( s) in the Nyquist contour);
∮𝑐 𝑊(𝑠)𝑑𝑠 = ∮𝑐 𝑑𝑟 + 𝑖 ∮𝑐 𝑑𝜃 (7) ∠𝐿 (−𝑗𝜔) = −∠𝐿 (−𝑗𝜔) (9)

The first term is zero, since r has the same value at the start and Hence the plot of L(s) when s moves on the boundary of the
end of the contour. If 𝜃 = 𝜃1 at the start of the contour and 𝜃 = Nyquist contour is just the polar plot + its symmetric plot about
𝜃2 at the end of the contour, then the real axis. The requirement for the open-loop condition of the
Nyquist on Bode plot have to be stable first, then in order for the
𝜃2 − 𝜃1 = −2𝜋 𝑖(𝑁𝑧 − 𝑁𝑝 ) (8) stable condition of the closed-loop, the Nyquist plot of L(s) should
not encircle the -1 point. In other words, the value of the |L(j ω)|
Nz is the number of zeros, and Np is the number of poles, taking have to be less than 1 whenever ∠L(j ω) = 180 (|L(j ω)| < 1). So
into account their multiplicity. This is the Principle of the Argu- the indicators in the Bode plot, that the magnitude plot should be
ment, and we can determine the difference between the number of less than 0 dB line if/when the phase plot crosses the −180 ◦ line.
zeros and poles of a transfer function by determining the number Note that the condition that already discussed above is valid only
of rotations that the transfer function angle makes, in the clock- if the stable condition for the open loop is approached. In another
wise (negative angle) direction, about the origin as the frequency case (including non-minimum phase zeros) it is strongly recom-
varies from −∞ 𝑡𝑜 + ∞ [1]. mended to double check any conclusion on closed-loop stability
by using other methods, for example by using Nyquist and root
3. Conventional Nyquist Design locus method.

R(s) C(s)
In this research, the use of the Nyquist stability criteria method
+ G(s)
-
where a plant that cannot be properly characterized can be handled.
Stability conditions can be seen from the results of the Frequency
response (performance in the Frequency domain) and the response
in the time domain [1]. Compared to the root placement method, H(s)
frequency response methods are easier and can be made accurate
with the availability of a sine generator and expand low frequen- Fig. 5: Close loop transfer function [7–8]
cies which are important in a system in accordance with the con- A system showing unstable or poorly damped oscillations may be
cept of this research. damped through the feedback of a stabilizer H(s) at a control loop
The Nyquist plot for a transfer function G(s) is the plotting of described by the transfer function G(s), as shown in Fig. 5. The
G(jω) in the complex plane (G(s)-plane) while is varied from −∞ open loop transfer function (OLTF) F(s) and the closed loop trans-
to +∞ . For linear dynamic systems with real coefficients, half of fer function (CLTF) T(s) are defined in (10) [7–8].
NP-related to ω varying from 0 to +∞ is a mirror image in the real
axis of the other half related to ω varying from 0 to −∞. Therefore,
International Journal of Engineering & Technology 25

𝐺(𝑠) A. Modelling of amplifier

𝐹(𝑠) = 𝐺(𝑆) 𝐻(𝑠) , 𝑇(𝑠) = (10)
1+𝐺(𝑠)𝐻(𝑠)
The transfer function G(s) of an amplifier is given in.
The Nyquist criterion states that the system will be stable if the
plane after the right of the G (s) H (s) curve does not cover the KA
Ga (s) = (13)
point (-1.0). The level of system stability can be measured by Gain 1+TA s
Margin (GM) and Phase Margin (PM).
The transfer function of amplifiers with 2 parameters is amplifiers
𝐺𝑎𝑖𝑛 𝑀𝑎𝑟𝑔𝑖𝑛 (𝐺𝑀) =
1
= 20 log10 𝑎 (𝑑𝐵) (11) gain constant (𝐾𝐴 ) and amplifiers time constant (𝑇𝐴 ). Amplifiers
𝑎 gain constant have a range value of 10.0000 until 40.000 while
amplifiers time constant have a range value of 0.0200 sec until
𝑃ℎ𝑎𝑠𝑒 𝑀𝑎𝑟𝑔𝑖𝑛 (𝐺𝑀) = −180 + 𝜃 (12) 0.1000 sec.
On a stable system, GM and PM values are always positive. The B. Modelling of exciter
greater the GM and PM values, the more stable the system is. The
Nyquist diagram is used to predict the stability and performance of The transfer function G(s) of the exciter is given in.
a closed-loop system by observing the behavior of the open loop.
In describing a Nyquist diagram, you must pay attention to the 𝐾𝐸
𝐺𝐸 (𝑠) = (14)
positive and negative frequencies (from zero to infinity).. 1+𝑇𝐸 𝑠

The transfer function of exciter with 2 parameters is exciter gain

constant (𝐾𝐸 ) and exciter time constant (TE )). Exciter gain con-
stant have a range value of 1.0000 until 10.000 while exciter time
constant have a range value of 0.4000 sec until 1.0000 sec.

C. Modelling of generator

The transfer function G(s) of the generator is given in.

𝐾𝐺
𝐺𝐺 (𝑠) = (15)
1+𝑇𝐺 𝑠
Fig. 6.. The Nyquist track system is stable [10].

To see the transient stability in a system that is by looking at the The transfer function of generator with 2 parameters is generator
"optimum" transient response if: [11] gain constant (𝐾𝐺 ) and generator time constant (𝑇𝐺 ). Generator
gain constant have a range value of 0.7000 until 1.000 while gen-
• The Phase margin of 30ᵒ to 60ᵒ
erator time constant have a range value of 1.0000 sec until 2.0000
• Margin gain> 6 dB
sec.
For the minimum phase system, the 300-600 phase margin means
that the slope of the Bode G (JW) curve at ω_gco must be more
D. Modelling of sensor
sloping than -40dB / dec (ie -20dB / dec) to be stable. If the slope
reaches -60 dB / dec, the system is almost certainly unstable. The transfer function G(s) of sensor is given in.
While the phase system is minimum stable when the gain margin
and the phase margin are positive. 𝐾𝑅
𝐻𝑆 (𝑠) = (16)
𝑇𝑅 𝑠+1
3.1. Modelling of Synchronous Generator
The transfer function of sensor with 2 parameters is sensor gain
This research begins by modeling 3 synchronous generators. Each constant (𝐾𝑅 ) and generator time constant (𝑇𝑅 ). Sensor gain con-
generator has the same rating: stant have a range value of 0.9000 until 1.1000 while generator
Spesification of generator time constant have a range value of 0.0010 sec until 0.0600 sec.
Type :TC30A
Rated output :2250 kW While electrical system modeling is a full single line diagram of a
Output voltage :6,6 kV system that includes generators, transformers and some of the
Frequency :50 Hz loads installed on the system in fig.8. There are a number of varia-
Current :237 A tions in the load including essential loads and non-essential loads.
Phase :3 ø In addition, the condition when the system is in a normal load has
Gas turbine :M1T-13A a power of 6467.82 kVA. While in peak load conditions it has a
Generator :NTAKL-DCK power of 7000.66 kVA. The following is a single line diagram
Synchronous Generator modeling includes amplifiers modeling, image of a system with 3 generators which are paralleled with
exciter modeling, generator modeling, and sensors modeling. The varying loads using ETAP 12.6 software.
form of a diagram block at Synchronous Generator modeling as
shown in Fig 7. Synchronous Generator modeling is performed
using of linear differential equations and Laplace transform.
Vref(s Ve(s) VR(s) Vf(s) Vt(s)
) +
GA(s) GE(s) Gg(s)
-

Amplifier Eksitasi Generator

Vs(s)
H(s)
Fig. 8: Single line diagram of the system
Fig. 7: Block diagram synchronous generator modelling [11]
26 International Journal of Engineering & Technology

4. Synchronous Generator Simulations

Transient stability analysis on 3 synchronous generators using the
Nyquist plot, and using the help of ETAP 12.6 software and also
MATLAB software.

4.1. Guidelines for Nyquist plot

The proposed method is an extension and improvement of the

conventional Nyquist design which is reviewed in this section.
The design based on the conventional Nyquist Plot (NP) is well-
known and used in different areas of control engineering [7–8].
Its main objective is to assess the stability of a system with feed-
back based on the open loop frequency response. The closed-loop Fig. 9: Nyquist diagram
transfer function (CLTF) (10) was selected to present the first
result. Some modifications were made to make the power plants. In accordance with the Nyquist diagram in fig.9, the curve G (jω)
The transfer function of equation (13) until (16) substitution into a is further away from the 1-point + j0 so that the system is said to
block diagram at fig.5 Then determine equation by transfer func- be unstable. The closeness of the G (jw) curve to the point -1 + j0
tion is given in. is a measure of the stability limit: phase margin and gain margin.

𝐾𝐴 +𝐾𝐸 +𝐾𝐺 +𝐾𝑅 4.2. Transient stability

𝐺𝐺 (𝑠)𝐻𝑆 (𝑠) = (17)
𝑆(1+𝜏𝐴 )(1+𝜏𝐸 )(1+𝜏𝐺 )(1+𝜏𝑅 )
For transient analysis of synchronous generators, it is simulated
Table 1: Parameter of gain constant and time constant using 3 synchronous generators. To do a parallel generator, sever-
Parameters Description Value al factors must be considered including the same frequency, the
𝐾𝐴 Amplifiers gain constant 1 same voltage, the same phase angle. The frequency of the electric
power system can be regulated by regulating the active power
𝐾𝐸 Exciter gain constant 1
generated by the generator. Active power regulation is closely
𝐾𝐺 Generator gain constant 1
related to the increase in the amount of fuel used to increase active
𝐾𝑅 Sensor gain constant 1.5 power. This fuel arrangement is done using a governor. So that it
𝑇𝐴 Amplifiers time constant 0.1 can be seen that the governor settings for each generator are as
𝑇𝐸 Exciter time constant 1.2 follows:
𝑇𝐺 Generator time constant 0.2 Ga Speed droop 4,5 %
𝑇𝑅 Sensor time constant 0.02 Gb Isochronous
The Parameter of gain constant and time constant are presented in
Gc Isochronous
Table 1 and can be reproduced by the formulas:
4.5
𝐺𝐺 (𝑠)𝐻𝑆 (𝑠) = 4 3 2
(0.00048𝑠 +0.0316𝑠 +0.41𝑠 +1.52𝑠+1) 4.2.1. Change of Source

The form above can be obtained value of phase margin and gain This simulation is based on case studies that occur in Pertamina
margin: EP where synchronization of Gb and Gc will maintain a frequency
Gain Margin 1.3045 of 50 Hz but when Ga is synchronized with the droop speed set-
ting of 4.5% with an initial frequency of 52.25 Hz it will decrease
GM frequency 6.9354
by 47.75 Hz. When Ga synchronizes with Gb and Gc in the 0.1-
Phase Margin 9.4647
second interval, all generators are still safe and synchronization is
PM frequency 6.0390 achieved. However, in the 0.18 interval, Ga is out of sync with Gb
Delay Margin 0.0274 and Gc and detects a decrease in frequency of 47 Hz where the
DM frequency 6.0390 frequency value has exceeded the predetermined setting limit so
In accordance with the above results, the Gain Margin is obtained that Gb and Gc with isochronous settings cannot maintain the
at 1.3045 dB (below 6.0000 dB) while the Phase Margin is ob- nominal frequency resulting in loss synchronization.
tained at 9,4647 ° (PM limit between 30,000 ° to 60,000 °).
When Ga will synchronize in the interval of 0.1 seconds, the gen-
So it can be concluded that the margin gain is below 6 dB and the erator is still safe, but in the interval of 0.18, it detaches by detect-
phase margin has a value below 30,000 ° does not cover the limit ing a frequency decrease of 47 Hz. Ga and Gb which is set to
which becomes the phase margin range itself so that the system isochronous cannot maintain its nominal frequency which results
can be said to be unstable. Plot a Nyquist diagram by using the in the generator blackout. The following is a comparison of field
form (17) as follows: data with data that has been simulated using ETAP 12.6 software.
Table 2: The result of a simulation load flow
Loading
Gen
MW Mvar
Data 1,477 1,018
Gen b
Simulation 1,408 0,787
Data 1,401 914
Gen c
Simulation 1,408 0,787
In table 2 shows, when loading before a disturbance occurs, the
difference in field data with simulation data is very small so that
the simulation can be done according to what happened in the
field. Each generator has the same generation power due to the
same governor setting, which is the isochronous governor setting.
International Journal of Engineering & Technology 27

The source changes in this system are determined by using the carried out using the Nyquist method implemented in MATLAB
governor settings of each generator and removing one of the and ETAP 12.6 software. The system stability is investigated
generator sources as seen from the transient response. Consists of through the governor settings and the transient response in ETAP
2 conditions where each generator has a different governor setting. 12.6 software. The transient response shows that the oscillations
are increasingly irregular and over a long period of time so that the
system is unstable. Besides that, it can be seen from the Gain
Margin value of 1.3045 dB while the Phase Margin is obtained at
9,4647 °. So, it can be said that the system is unstable, and the
transient response tends to be minimum.

Acknowledgement
The authors would like to thank Pertamina EP for contributing in
providing the case studies and the data needed to support this re-
(a) (b) search. Also, many thanks to the Energy Conversion and Power
system Laboratory, Electrical Engineering Department, State
Fig. 10: (a). Gb and Gc in isochronous settings, (b). When it doesn't come
Polytechnic of Malang that has provided many insights. This re-
off with GB in the droop and Gc speed settings in the isochronous setting
search was supported by State Polytechnic of Malang.
The condition of the different governor settings can be seen in
Fig.10, that when synchronization occurs between 3 generators References
experiencing poor oscillation and the system cannot return to its
steady state condition and the frequency of the lowers is 45-47 Hz.
[1] Kundur,Praba, “Power System Stability And Controll," Washington
Dc, Mc Graw Hill, 1994.
4.2.2. Change of Load [2] Jun Zhou, Qiufang Zhang, Huimin Qian, “ s-Domain Generalized
Nyquist Criterion For Double Loop Sampled-Data Systems", in
In this change simulation, the system is set based on varying load 2015 IEEE International Conference on Chinese Automation Con-
changes with the aim of knowing the effect of the generator based gress (CAC), pp. 139-144, 2015.
on the load power supplied. Then the simulation results are ob- [3] K. Ogata, Modern Control Engineering, Prentice Hall, 2001.
tained as follows: [4] S. Gomes, Guimarães C.H.C, Martins N, Taranto G.N., “Damped
Nyquist Plot for a pole placement design of power system stabi-
lizers”, Electrical Power Research, pp.158-169, January 2018.
[5] Rygg Atle, Amin Mohammad, Molinas Marta, “Apparent imped-
ance analysis: A new method for power system stability analysis”,
in IEEE Journal of Emerging and Selected Topics in Power Elec-
tronic, Vol.5, no.4, pp.1474-1486, 2017.
[6] Busetti Hadrien, Sohail Khan, “Transient Stability In Oscillating
Multi-Machine System Using Lyapunov Vectors", in IEEE Trans-
action On Power System, Vol.33, no.2, pp. 2078-2086, 2018.
[7] Supegina Fina, “Dasar Sistem Kontrol”, in Pusat Pengembangan
(a) (b) Bahan Ajar, Universitas Mercu Buana, Vol 12 , 2016.
[8] B.C. Kuo, Automatic Control Systems, 7th ed., John Willey & Sons,
New York, NY, 1995
[9] Razzaghi Reza, Davarpanah Mahdi, And Majid Sanaye-Pasand,
“A Novel Protective Scheme To Protect Small-Scale Synchronous
Generators Against Transient Instability", in IEEE Transactions on
Industrial Electronics, Vol 60, no.4, pp. 1659-1667, April 2013.
[10] Qiufang zhang, Jun Zhou’, Xinbau LU, Huimin Qian, ”Stability
Analysis Via Contraposition Nyquist Approach For Sampled-Data
(c) System With Auxiliary Continuos-Time Feed Back" in IEEE Inter-
national Conference on The 27th Chinese Control and Decision
Fig. 11: (a) losses A and B Generators (2.8 MW Load) (b). Loose A and B
Conference (CCDC), July 2015.
generators (2.6 MW load) (c). Loose A and B generators (1.9 MW load)
[11] Turner Robert, Walton Simon, And Duke Richard, “A Case Study
On Aplication Of The Nyquist Stability Criterion As Applied To
With varying load conditions in fig.11, the transients shown are Interconnected Loads And Sources On Grids" in IEEE Transaction
On Industrial Electronics, Vol.60, no.7, pp.2740-2749, July 2013.
not optimum and are still experiencing continuous oscillation so
that further handling needs to be done so that the oscillations do
not experience too long which results in the system becoming
worse. The system will experience continuous oscillation until it
cannot return to its steady state position. If left unchecked it will
experience a synchronization failure. Then it needs to be analyzed
from the governor's settings. The solution of this research is by
modeling more than one generator using the equation of the trans-
fer function, it can be seen transient stability with different con-
trols. In addition, pay attention to the effect of the angle on the
rotor which is related to the regulation of active power (W) and
reactive power (VAR) or fuel regulation on each generator.

5. Conclusion
This paper aims to analyze the transient stability of three 2.25 MW
6.6 kV generators which are operated in parallel. The analysis is