Number 1

ISSN: 2085-6350

PROCEEDINGS OF
CONFERENCE ON
INFORMATION TECHNOLOGY
AND ELECTRICAL ENGINEERING

INTERNATIONAL SESSION
Electrical Power Systems

DEPARTMENT OF ELECTRICAL ENGINEERING

FACULTY OF ENGINEERING
GADJAH MADA UNIVERSITY

Conference on Information Technology and Electrical Engineering (CITEE)

Organizer
Steering Commitee
Adhi Susanto (UGM)
Hamzah Berahim (UGM)
Thomas Sri Widodo (UGM)
Dadang Gunawan (UI)
Heri Mauridi (ITS)
Yanuarsyah Harun (ITB)
Anto Satrio Nugroho (BPPT)
Son Kuswadi (PENS)
Advisory Board
Tumiran (UGM)
Lukito Edi Nugroho (UGM)
Anto Satrio Nugroho (BPPT)
Son Kuswadi (PENS)
General Chair
Bambang Sutopo
Organizing Chairs
Risanuri Hidayat
Sri Suning Kusumawardhani
Ridi Ferdiana
Adha Imam Cahyadi
Budi Setiyanto
Program Chairs
Prapto Nugroho
Agus Bejo
Cuk Supriyadi Ali Nandar (BPPT)
Yusuf Susilo Wijoyo
Publication Chair
Enas Dhuhri K
Finance Chairs
Eny Sukani Rahayu
Maun Budiyanto
Roni Irnawan
Secretariats
Astria Nur Irfansyah
Lilik Suyanti

YOGYAKARTA, AUGUST 4, 2009

Conference on Information Technology and Electrical Engineering (CITEE) 2009

FOREWORD
First of all, praise to Almighty God, for blessing us with healthy and ability to come here, in the
Conference of Information and Electrical Engineering 2009 (CITEE 2009). If there is some noticeable wisdoms
and knowledge must come from Him.
I would like to say thank you to all of the writers, who come here enthusiastically to share experiences and
knowledge. Without your contribution, this conference will not has a meaning.
I also would like to say thank you to Prof. Dadang Gunawan from Electrical Engineering, University of
Indonesia (UI), Prof. Yanuarsyah Haroen from Electrical Engineering and Informatics School, Bandung Institute
of Technology, ITB, Prof. Mauridhi Hery Purnomo from Electrical Engineering Department, Surabaya Institute
of Technology (ITS). And also Prof. Takashi Hiyama from Kamamoto University, Japan, Thank you for your
participation and contribution as keynote speakers in this conference.
This conference is the first annual conference held by Electrical Engineering Department, Gadjah Mada
University. We hope, in the future, it becomes a conference of academics and industries researchers in the field of
Information Technology and Electrical Engineering around the world. We confine that if we can combine these
two fields of sciences, it would make a greater impact on human life quality.
According to our data, there are 140 writers gather here to present their papers. They will present 122 titles
of papers. There are 47 papers in the field of Electrical Power Systems, 53 papers in the area of Systems, Signals
and Circuits, and 22 papers in Information Technology. Most of these papers are from universities researchers.
We hope, the result of the proceedings of this conference can be used as reference for the academic and
practitioner researchers to gain
At last, I would like to say thank you to all of the committee members, who worked hard to prepare this
conference. Special thanks to Electrical Engineering Department, Gadjah Mada University, of supporting on
facilities and funds.
Thank you and enjoy the conference, CITEE 2009, and the city, Yogyakarta
August, 4Th, 2009

Bambang Sutopo

Electrical Engineering Dept., Fac. of Engineering, GMU

Proceedings of CITEE 2009

Number 1

ISSN: 2085-6350

Table of Contents
Organizer
Foreword
Table of Contents
Schedule

ii
iii
v
vii

KEYNOTE
Social Intelligent on Humanoid Robot: Understanding Indonesian Text Case Study
Mauridhi Hery Purnomo (Electrical Engineering Department, ITS, Indonesia)

Signal Processing: Video Compression Techniques

Dadang Gunawan (Electrical Engineering Department, University of Indonesia)

Intelligent Systems Application to Power Systems

Prof. Takashi Hiyama (Kumamoto University, Japan)

INTERNATIONAL SESSION: Electrical Power Systems

Control System Integrated Starter - DC Motor Couple Three Phase Induction Motor for Automotive Applications
Zulkarnain Lubis, Ahmed N. Abdalla, Samsi bin MD said .Mortaza bin Mohamed

Transient Stability of SMIB: a Case Study

Adelhard Beni Rehiara

12

Look-Up Table of Fuzzy Rule SURAM with AVR ATmega128

Zakarias Situmorang

17

Optimal Capacitor Bank Location in the Primary Feeder with Typical Flat Load
Hermagasantos Zein

25

Dual Parallel Power Conversion Converter Supplied by Photovoltaic for Base Transceiver Station (BTS) Power Supply
Kartono Wijayanto, Yanuarsyah Haroen

30

Integrated Buck Converter and Wind Turbine Control System Medium Scale (100 W) for Optimization Wind Power and
Electricity Power
Ali Musyafa, Soedibjo, I Made Yulistiya Negara , Imam Robandi

36

Optimal Power Flow Analysis Using Genetic Algorithm in 500 KV Java Bali Interconnection System
Buyung Baskoro, Adi Soeprijanto, Ontoseno Penangsang

43

Transient Stability Assessment of Java Bali 500 KV Multi Machine Electrical Power System Using Committee Neural
Network
Eko Prasetyo, Boy Sandra, Adi Soeprijanto

49

Overcurrent Protection Coordination Due to Liquid Starter Effect on Large Induction Motor
Dimas Anton Asfani, Nalendra Permana

57

Analysis of 20 KV PLN Relay Protection Typical Setting in Industrial Customer

Dimas Anton Asfani, Iman Kurniawan, Adi Soeprijanto

62

The Electrical Energy Calculation Based Upon the Voltage Measurement of a High Speed and Magnitude of Multiple
Impulse Currents Produced by an Impulse Generator Implemented to ZnO Block
Haryono. T, Sirait K.T., Tumiran, Hamzah Berahim

69

Study of Return Voltage Transient in Low Voltage ZnO Arrester Type OBO Bettermann V-20 C/1
Nurcahyanto, T. Haryono, Suharyanto

73

The Design of Digital Overcurrent Relay with IEC 60255 Time Curve Characteristic Based on an ATmega16
Microcontroller
Agni Sinatria Putra, Tiyono, Astria Nur Irfansyah

77

Conference on Information Technology and Electrical Engineering (CITEE)

Conference on Information Technology and Electrical Engineering (CITEE)

SCHEDULE
Tuesday, August 4, 2009
07.30 08.00:
08.00 08.15:

Registration
Opening
1. Welcome speech by conference chairman
2. Speech by GMUs Rector
08.15 09.20:
PLENARY SESSION
Prof. Takashi Hiyama (Kumamoto University, Japan): Intelligent Systems Application to
Power Systems
Prof. Dr. Mauridhi Hery Purnomo (Electrical Engineering Department,ITS, Indonesia):
Social Intelligent on Humanoid Robot: Understanding Indonesian Text Case Study
Prof. Dr. Dadang Gunawan (Electrical Engineering Department, University of
Indonesia): Signal Processing: Video Compression Techniques
Prof. Dr. Yanuarsyah Haroen (Electrical Engineering and Informatics School, ITB,
Indonesia): Teknologi Sistem Penggerak dalam WahanaTransportasi Elektrik
09.20 09.30: Break
PARALLEL SESSION
INTERNATIONAL SESSION (Room 1, 2)
Room: 1
Time

Group

Country/City

09.30 09.45

Malaysia

09.45 10.00
10.00 10.15
10.15 10.30
10.30 10.45
10.45 11.00
11.00 11.15
11.15 11.30
11.30 11.45
11.45 12.00
12.00 13.00
13.00 13.45
13.15 13.30

P
P
P

Papua
Medan
Bandung

P
P
P
P
P

Bandung
Surabaya
Surabaya
Surabaya
Yogyakarta

P
P

Surabaya
Surabaya

13.30 13.45

Surabaya

13.45 14.00
14.00 14.15
14.15 14.30
14.30 14.45
14.45 15.00
15.00 15.15
15.15 15.30

I
I

Surabaya
Surabaya

I
I
P
P

Yogyakarta
Surakarta
Yogyakarta
Yogyakarta

Author(s) or Presenter(s)
Zulkarnain Lubis, Ahmed N. Abdalla, Samsi bin MD said .Mortaza bin
Mohamed
Adelhard Beni Rehiara
Zakarias Situmorang
Kartono Wijayanto, Yanuarsyah Haroen
Coffee Break
Hermagasantos Zein
Ali Musyafa, Soedibjo, I Made Yulistiya Negara , Imam Robandi
Buyung Baskoro, Adi Soeprijanto, Ontoseno Penangsang
Eko Prasetyo, Boy Sandra, Adi Soeprijanto
T. Haryono, Sirait K.T., Tumiran, Hamzah Berahim
Lunch Break
Dimas Anton Asfani, Nalendra Permana
Dimas Anton Asfani, Iman Kurniawan, Adi Soeprijanto
F.X. Ferdinandus, Gunawan, Tri Kurniawan Wijaya, Novita Angelina
Sugianto
Arya Tandy Hermawan, Gunawan, Tri Kurniawan Wijaya
Herman Budianto, Gunawan, Tri Kurniawan Wijaya, Eva Paulina Tjendra
Coffee Break
Bambang Soelistijanto
Munifah, Lukito Edi Nugroho, Paulus Insap Santosa
Nurcahyanto, T. Haryono, Suharyanto.
Agni Sinatria Putra, Tiyono, Astria Nur Irfansyah

Notes:
1. P: Electrical Power Systems; S: Signals, Systems, and Circuits; I: Information Technology
2. Paper titles are listed in Table of Contents

Department of Electrical Engineering, Faculty of Engineering, Gadjah Mada University

Conference on Information Technology and Electrical Engineering (CITEE)

Room: 2
Time
09.30 09.45
09.45 10.00
10.00 10.15
10.15 10.30
10.30 10.45
10.45 11.00
11.00 11.15
11.15 11.30
11.30 11.45
11.45 12.00
12.00 13.00
13.00 13.45
13.15 13.30
13.30 13.45
13.45 14.00
14.00 14.15
14.15 14.30
14.30 14.45
14.45 15.00
15.00 15.15
15.15 15.30

Group
S
S
S
S

Country/City
INDIA
Jakarta
Jakarta
Jakarta

I
I
I
S
S

Yogyakarta
Yogyakarta
Lampung
Semarang
Semarang

S
S
S
S
S

Yogyakarta
Yogyakarta
Yogyakarta
Yogyakarta
Yogyakarta

S
S
S
S

Yogyakarta
Yogyakarta
Yogyakarta
Yogyakarta

Author(s) or Presenter(s)
Ms.M.Thanuja, Mrs. K. SreeGowri
A. Suhartomo
Riandini, Mera Kartika Delimayanti, Donny Danudirdjo
Purnomo Sidi Priambodo, Harry Sudibyo and Gunawan Wibisono
Coffee Break
Arwin Datumaya Wahyudi Sumari, Adang Suwandi Ahmad
Arwin Datumaya Wahyudi Sumari, Adang Suwandi Ahmad
Sumadi, S; Kurniawan, E.
Florentinus Budi Setiawan
Siswandari N, Adhi Susanto, Zainal Muttaqin
Lunch Break
Thomas Sri Widodo, Maesadji Tjokronegore, D. Jekke Mamahit
Tarsisius Aris Sunantyo, Muhamad Iradat Achmad
Muhamad Iradat Achmad, Tarsisius Aris Sunantyo, Adhi Susanto
Usman Balugu, Ratnasari Nur Rohmah, Nurokhim
Okky Freeza Prana Ghita Daulay, Arwin Datumaya Wahyudi Sumari
Coffee Break
Sri Suning Kusumawardani and Bambang Sutopo
Risanuri Hidayat
Budi Setiyanto, Astria Nur Irfansyah, and Risanuri Hidayat
Budi Setiyanto, Mulyana, and Risanuri Hidayat

NATIONAL SESSION (Room 3, 4, 5, 6, 7)

Yogyakarta, August 4, 2009

Proceedings of CITEE, August 4, 2009

-1

Keynote

Social Intelligent on Humanoid Robot:

Understanding Indonesian Text
Case Study
Mauridhi Hery Purnomo
Electrical Engineering Department-Institut Teknologi Sepuluh November
Surabaya 60111, Indonesia
[email protected]
Abstract Social affective and emotion are required on
humanoid robot performance to make the robot be more
human. Social intelligent are the individual ability to manage
relationship with other agents and act wisely based on
previous learning experiences. Here, social intelligent is
intended to understand Indonesian text. How the computation
process, as well as affective interaction, emotion expression of
the humanoid robot to the human statement. This process is a
highly adaptive complex approximation, dependently on its
entire situation and environment.
Keywordssocial, affective, emotion, intelligent, computing
(key words)

I. INTRODUCTION
Social and interactive behaviors are necessary
requirements in wide implementation areas and contexts
where robots need to interact and collaborate with other
robots or humans. The nature of interactivity and social
behavior in robot and humans is a complex model.
An experimental robot platform KOBIE, which provides
a simulation tool for emotion expression system includes an
emotion engine was developed. The simulation tool
provides a visualization interface for the emotion engine
and expresses emotion through an avatar. The system can be
used in the development of cyber characters that use
emotions or in the development of an apparatus with
emotion in a ubiquitous environment [1].
To improve the understandability and friendliness in
human-computer interfaces and media contents, a
Multimodal Presentation Markup Language (MPML) is
developed. MPML is a simple script language to make
multi-modal presentation contents using animated
characters for presenters [2].
Other effort in the robot head which uses arm-type
antennae, eye-expression, and additional exaggerating parts
for dynamic emotional expression is also developed. The
robot head is developed for various and efficient emotional
expressions in the Human-Robot interaction field. The
concept design of the robot is an insect character [3].
In regard to artificial cognitive, iCub humanoid robot
systems is developed. The system is open-systems 53
degree-of-freedom cognitive humanoid robot, 94 cm tall,
the same size as a three year-old child. Able to crawl on all
fours and sit up, its hands will allow dexterous
manipulation, and its head and eyes are fully articulated. It

has visual, vestibular, auditory, and haptic sensory

capabilities [4].
An innovative integration of interactive group learning,
multimedia technology, and creativity used to enhance the
learning of basic psychological principles was created.
This system is based on current robotic ideology calling for
the creation of a PowerPoint robot of the humanoid type
that embodies the basic theories and concepts contained in a
standard psychological description of a human being [5].
Now days, not only visual and auditory information are
used in media and interface fields but also multi-modal
contents including documents such like texts. Thus, in this
paper, a part of result on emotion expression and
environment through understanding Indonesian text, as
affective interaction between a human and a robot is
explored.
The paper is organized as follows. In Section 2, the
general emotion and expression system on life-like agent is
presented. Section 3 describes the experimental on
Indonesian text classification. In Section 4, the preliminary
result in emotion classification of Indonesian article text is
discussed.
II.

EMOTION AND EXPRESSION ON LIFE-LIKE AGENT

Social Computing, Social Agent, and Life-likeness

Many Psychologists have studied a definition and
classification of emotions, therefore, so many classification
methods of emotions and expressions. However, we need to
choose categories of emotions that are suitable expressed by
robot, as well as the well-known Ekmans 6 basic emotion
expressions model can be used.
In social computing life-like characters are the key, and
the affective functions create believability. To articulate
synthetic emotions can be presented as; personalities, human
interactive behavior or presentation skills. The personalities;
by means of body movement, facial display, and the
coordination of the embodied conversational behavior of
multiple characters possibly including the user. Personality
is key to achieving life-likeness
Some Applications of Life-Like Character
Life-like characters are synthetic agents apparently living
on the screen of computers. Life-like character can be
implemented as virtual tutors and trainers in interactive
learning environments. On the web as an information expert,
presenter, communication partners, and enhancing the search
engine. The other application as actors for entertainment, in

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

Keynote

online communities and guidance systems as personal

representatives.
Early characterization of the emotional and believable
character was raised by Joseph Bates. He said, the portrayal
of emotions plays a key role in the aim to create believable
characters, one that provides the illusion of life, and thus
permits the audiences suspension of disbelief. In game and
animation, suspension of disbelief is very important, for
instance as: synthetic actors, non-player characters, and
embodied conversational agents.

Proceedings of CITEE, August 4, 2009

Figure 1 is an example how to train machine (agent) in

order responsive to the external and adequate response. The
case study is, Indonesian text classification. There are two
types of machine learning, supervised and unsupervised
learning.
Figure 2 show a block diagram process of embodied
conversational agent.
IV.

Emotion and personality are often seen as the affective

bases of believability, and sometimes the broader term social
is used to characterize life-likeness.

III.

CLASSIFICATION SYSTEM FOR INDONESIAN TEXT

Information growth, including texts are faster than

human ability, thus help system is quite necessary. For
instance as the following illustration;
y

The Recent study, which used Web searches in 75

different languages to sample the Web, determined
that there were over 11.5 billion (1012) Web pages
in the publicly indexable Web as of the end of
January 2005

As of March 2009, the indexable web contains at

least 25.21 billion pages

On July 25, 2008, Google software engineers Jesse

Alpert and Nissan Hajaj announced that Google
Search had discovered one trillion unique URLs

As of May 2009, over 109.5 million websites

operated

feature
extractor

..features..

test
input

language
dependent
NLP tools

feature
extractor

..features..

machine
learning

classifier
model

When I nearly walked on a blindworm and then

saw it crawl away disgust

When I was involved in a traffic accident fear

I do not help out enough at home guilt

Passing an exam I did not expect to pass joy

Failing an examination sadness

When, as an adult I have been caught lying or

behaving badly shame

The classification is divided into six (6) classes of

emotion: disgust, shame, anger, sadness, joy and fear.
Each class has 200 text files, data: as-is; DataNot:
pre-processing only handles not. Split ratio 0.5 shows
f-measure scores 0.59

label
language
dependent
NLP tools

The following sentences are sample of statements,

and some emotion expressions;
When a car is overtaking another and I am forced
to drive off the road anger

Based on some statements and the emotion expression as

mentioned above in Indonesian text, the preliminary
classification results are shown in the table 1, figure 3 and
figure 4.

(a) training phase

training
input

EMOTION CLASSIFICATION FROM INDONESIAN

ARTICLE TEXT

Pre-processing Steps
predicted
label

(b) prediction phase

Figure 1. Example of Indonesian Text Classification

Text Classification (TC) techniques usually ignore stopwords and case of input text. In pre-processing step, stopwords removal can be applied.
Stop-words such as not, in, which and exclamation
marks (!) usually do not affect categorization of text.

Text
Input

Knowledge
Base

Response
User
(Human)

Text-based
Conversational Agent

Information
Retrieval
Text
Classification
Text
Mining

Figure 2. Knowledge from Free (Unstructured) Text

The illustrations as mentioned above explain the

essential of help system especially in Indonesian text. We
have developed a system for understanding and classifying
Indonesian text, and the block diagram as shown in figure 1
and figure 2.

ISSN: 2085-6350

TABLE I. NAVE BAYES CLASSIFICATION INTO 4 CLASS

Free
Text

Classifica
tion ratio
(%)

Usual & original text

Text without stop

words

20

71.41

69.81

40

73.33

71.3

60

74.40

71.3

80

75.33

74.05

Accuracy (%)

Example:
-

Microsoft released Windows categorized as

news.

Microsoft has not released Windows yet still

categorized as news.

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

Input text is converted

classifications, as example:

to

-3

Keynote

lowercase

on

emotion

DataNot

I do like you. I do not like you!

70

I do not like you. I DO NOT LIKE YOU!!

65
60
F-Measure

Naive Bayes
70

55
50

65

Multinomial NB
45

NB

Precision

60
40

55

0,1

0,2

0,4

0,5

0,6

0,7

0,8

0,9

Rasio data

50

Data
DataNot

With pre processing

45

Figure 4. Results Recapitulation of Indonesian Text Classification

40
40

45

50

55

60

65

70

Recall

V.

Multinomial NB
70
65
60
55

REFERENCES

Data
DataNot

50

[1]

45
40

[2]
40

45

50

55

CONCLUDING REMARKS

The preliminary results of emotion expression and related

environment through Indonesian text are described. We
develop an Indonesian conversational agent system includes
an emotion expression engine, that will used in the game
engine. The use of emotion on Indonesian text is expected
to the improvement of expressiveness of understanding and
actions. The research are still underway, so many possibility
to improve our future works and making the system more
life-like.

non bayesian

Precision

0,3

60

65

70

Recall

[3]
Multinomial non bayesian
Figure 3. Emotion Classification of Indonesian Text
[4]
Data
65

[5]
60

C Park, J W Ryu, J Kim, S Kang, J Sohn, YJ Cho, Emotion

Expression and Environment Through Affective Interaction
Proceedings of the 17th World Congress The International Federation
of Automatic Control,Seoul, Korea, July 6-11, 2008 .
K Kushida, Y Nishimura, et al.Humanoid Robot Presentation
through Multimodal Presentation Markup Language MPML-HR
AAMAS05, Utrecht, Netherlands, July 25-29, 2005.
H Song and DS Kwon, Design of a Robot Head with Arm-type
Antennae for Emotional Expression, International Conference on
Control, Automation and Systems in COEX, Seoul, Korea Oct. 1720, 2007.
G Sandini, G Metta, and D Vernon,The iCub Cognitive Humanoid
Robot:An Open-System Research Platform for Enactive Cognition,
M. Lungarella et al. (Eds.): 50 Years of AI, Festschrift, LNAI 4850,
Springer-Verlag Berlin Heidelberg, pp. 359370, 2007.
James L. Anderson and Erin M. Applegarth, The Psychological
Robot: A New Tool for Learning, 3rd ed., International Journal of
Teaching and Learning in Higher Education 2007, Volume 19,
Number 3, 305-314

F-Measure

55

50

Multinomial NB

45

Naive Bayes
40
0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

Rasio data

Without pre processing

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

Keynote

Proceedings of CITEE, August 4, 2009

Signal Processing:
Video Compression Techniques
Dadang Gunawan
Electrical Engineering Department, University of Indonesia
In our information society, signal processing has been
created a significant effect. Signal processing can be found
everywhere: in home appliances, in Cell Phone, TVs,
Automobile, GPSs, Modem Scanner, and All kind of
Communication Systems and Electronic Devices. Modern
cell phones are indeed a most typical example within
these small wonders, voice, audio, image, video and
graphics are processed and enhanced based on decades of
media signal processing research.
Technological advancement in recent years has
proclaimed a new golden age for signal processing [1].
Many exciting directions, such as bioinformatics, human
language, networking, and security, are emerging from
traditional field of signal processing on raw information
content. The challenge in the new era is to transcend from
the conventional role of processing in low level, waveformlike signal to the new role of understanding and mining the
high-level, human-centric semantic signal and information.
Such a fundamental shift has already taken place in limited
areas of signal processing and is expected to become more
pervasive in coming years of research in more areas of
signal processing.
One of the huge applications of signal processing is
exploited as video compression. Nowadays, video
applications such as digital laser disc, electronic camera,
videophone and video conferencing systems, image and
interactive video tools on personal computers and
workstations, program delivery using cable and satellite,
and high-definition television (HDTV) are available for
visual communications. Many of these applications,
however, require the use of data compression because visual
signals require a large communication bandwidth for
transmission and a large amounts of computer memory for
storage [2][3]. In order to make the handling of visual
signals cost effective it is important that their bandwidth be
compressed as much as possible. Fortunately, visual signals
contain a large a mount of statistically and psychovisually
redundant information [4]. By removing this unnecessary
information, the amount of data necessary to adequately
represent an image can be reduced.
The removal of unnecessary information generally can
be achieved by using either statistical compression
techniques or psychovisual compression techniques. Both
techniques result in a loss information, but in the former the
loss may be recovered by signal processing such as filtering
and inter or intra-polation. In the later, information is in fact
discarded, but in way that is not perceptible to a human
observer. The later techniques offer much greater levels of
ISSN: 2085-6350

compression but it is no longer possible to perfectly

reconstructed the original image [4]. While the aim in
psychovisual coding is to keep these differences at an
imperceptible level, psychovisual compression inevitably
involves a tradeoff between the quality of the reconstructed
image and the compression rate achieved. This tradeoff can
often be assessed using mathematical criteria, although a
better assessment is in general provided by human observer.
The applications of image data compression, in general,
are primarily in the transmission and storage of information.
In transmission, applications such as broadcast television,
teleconferencing, videophone, computer-communication,
remote sensing via satellite or aircraft, etc., require the
compression techniques to be constrained by the need
For the real time compression and on-line consideration
which tends to severely limit the size and hardware
complexity. In storage applications such as medical images,
educational and business documents, etc., the requirements
are less stringent because much of the compression
processing can be done off-line. However, he
decompression or retrieval should still be quick and
efficient to minimize the response time [5].
All images of interest usually contain a considerable
amount of statistically and subjectively superfluous
information [6]. A statistical image compression technique
exploits statistical redundancies in the information in the
image. This technique reduces the amount of data to be
transmitted or to be stored in an image without any
information being lost. The alternative is to discard the
subjective redundancies in an image, which leads to
psychovisual image compression. These psychovisual
techniques rely on properties of the Human Visual
characteristic system (HVS) to be determined which
features will not be noticed by human observer.
There are numerous way to achieve compression in
statistical image compression techniques such as Pulse
Code Modulation (PCM), Differential PCM (DPCM),
Predictive Coding, Transform Coding, Pyramid Coding and
Subband Coding, as well as Psychovisual Coding
techniques. Statistical compression techniques all use a
form of amplitude quantization in their algorithms to
improve compression performance. Simple quantization
alone, however, is not the most efficient or flexible
techniques to combine with a statistical compression
algorithm [7]. The combination of quantization and
psychophysics, on the other hand has the potential to
remove most subjectively redundant information efficiently

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

Keynote

from an image, a process which is based on the actual

behavior of the HVS. Furthermore, subband coding and
pyramid coding schemes can be combined with visual
psychophysics-based compression techniques, since both of
these statistical schemes break the original image data down
into separate frequency bands. A process that similar to the
bandpass filter characteristics of the HVS and can be used
to quantized the information in each band depending on the
relative frequency band.
Transform coding is able to achieve optimum statistical
compression ratios, especially the Discrete Cosine
Transform (DCT). Much research has been performed in
combining the DCT transform coding and visual
psychophysics-based compression techniques [8][9][10]
resulting in a higher compression ratio and good
reconstruction of the original image.
Image compression techniques mentioned above,
involve spatial correlations in single frames where
redundancies are exploited either statistically or
subjectively, are known as intraframe coding techniques.
Interframe coding techniques known video compression, by
contrast, attempt to exploit the redundancies produced by
temporal correlation as well as spatial correlations in
successive video signals. These techniques hold the promise
of significantly greater reduction in the data required to
transmit the video signal as compared to interframe coding.
The simplest interframe coding technique is called
conditional replenishment [11][12][13]. This technique
bases the coding scheme on the previous frame and is also
often called predictive coding. In the conditional
replenishment technique, only pixels the values of which
have changed significantly since the last frame, as
compared to a certain threshold, are transmitted. Another
technique, which still uses predictive coding from previous
frame, is adaptive intra-inter-frame prediction [14]. In this
technique, interframe prediction is used for scenes in
images where there is little motion, while intraframe
prediction is used for areas where this is much motion. The
switching between intra- and inter-frame prediction or a
combination of both, is usually controlled by the signal
changes of previously transmitted pixels so that no
overhead control information need to sent. The prediction
error can be quantized and transmitted for every pixel or can
be thresholded into predictable and non-predictable pixels
[15].
Adaptive prediction displacement of a moving object
which is based on information obtained from successive
frames is known as Motion Compensation. This scheme
was studied by [16] and [17] by measuring small
displacements based on very simple model of moving
objects in a stationary background scene for segmentation
purposes. A later refinement developed by [18][19][20] led
to one set of techniques known as Pel Recursive Motion
Compensation, which recursively adjusts the translational
estimates at every pixel or every small block of pixels. [21]
developed another technique known Block Matching
Motion Estimation. This technique estimates the location of
a block of pixels in the current frame by using a search in a

-5

confined window defined in the previous frame. Location of

the block results in the displacement vector for that block.
Different search methods have been proposed to avoid an
exhaustive search [22][23][24][25].
In order to produce a higher compression ratio,
transform coding has been applied to video coding, and can
be carried out as a three-dimensional transform [26] or in an
interframe coding scheme [27][28]. In the latter case motion
compensation can be performed in either the spatial domain
or the frequency domain. Transform coding can also be
combined with predictive coding so that the transform
coefficients from intraframe transformations of the previous
frame can be used to predict the transform coefficients of
the current frame [29]. CCITT H.261 Recommendations
[30], JPEG standard [31] and the MPEG draft [32], are also
DCT transform based and intra-inter-frame adaptive with
optional motion compensation. Their schemes result in a
blocking effect for low bit rates. Wavelet transform coding
can
effectively
eliminate
this
blocking
effect
[33][34][35][36][37] because the wavelet bases in adjacent
subbands overlap one another. Another advantage of
wavelet transform coding is that it is very similar to
subband coding. Wavelet transform combined by
psychovisual coding resulted a very good performance in
term of compression ratio and reconstructed images [4].
Since then, the DCT is replaced to the WT in order to
achieve high compression algorithms and good quality
reconstructed images, and has been adopted to be standard.
These standards are ITU standard for H-261, H-263, H-264;
ISO/IEC for JPEG, JPEG-2000, MPEG-2, MPEG-4, and
MPEG-7.
However, there is inevitably space for
improvements and extension within this area of research,
such as a hybrid system by using combining transform
method and Fuzzy, Neural Network, etc. For instance, the
TEMATICS Team has been developed some algorithm and
practical for analysis and modeling of video sequences;
sparse representations, compression and interaction with
indexing; Distributed source & Joint Source-Channel
Coding, etc [38].
References:
[1]
Li Deng, Embracing A new Golden Age of Signal
Processing, IEEE Signal Processing, Jan., 2009.
[2]
Dadang Gunawan, Interframe Coding and Wavelet
Transform, Journal IEICE, Vol. 1, No 1, pp. 22 37, Oct.,
1999.
[3]
Dadang Gunawan, From Image to Video Compression,
Jurnal Teknologi, Vol. IX, No. 2, Sep., 1995.
[4]
Dadang Gunawan & D.T. Nguyen, Psychovisual Image
Coding using Wavelet Transform, Australian Journal of
Intelligent Information Processing Systems, Autumn
Issues, Vol. 2, No. 1, Mar.,1995.
[5]
A. K. Jain, Image Data Compression, Proceeding of the
IEEE, Vol. 69., pp. 349 389, Mar., 1981.
[6]
Arun N Netravali & Barry G Haskell Digital Pictures :
Representation and Compression, Plenum Press, new
York, 1988.
[7]
David L McLarent, Video and Image Coding for
broadband ISDN, Ph.D. Thesis, University of Tasmania,
Australia, 1992.

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

6
[8]
[9]
[10]
[11]
[12]
[13]
[14]

[15]
[16]

[17]
[18]
[19]

[20]

[21]

[22]

Keynote
K. N. Ngan, K. S. Leong & H. Singh, Adaptive Cosine
Transform Coding of Images in perceptual Domain, IEEE
Transaction ASSP, Vo. 37, pp. 1743 1750, Nov., 1989.
B Chitprasert & K. R. Rao, Human Visual Weighted
progressive Image Transmission, IEEE Transaction on
Communication, Vol. 38, pp. 1040 1044, Jul. 1990.
D. L. McLaren & D. T. Nguyen, The Removal Subjective
redundancy fro DCT Coded Images, IEE Proceeding
Part I, Vol. 138, pp. 345 350, Oct. 1991.
F. W. Mounts, A Video Coding System with Conditional
Picture-Element Replenishment, The Bell System
Technical Journal, Vol. 48, pp. 2545 2554, Sep. 1969.
J. C. Candy, Transmitting television as Clusters of Frame
to frame Differences, The Bell System Technical Journal,
Vol. 50, pp. 1889 1917, Aug. 1971.
[B. G. Haskel, F. W. Mount and C. Candy, Interframe
Coding of Videotelephone Pictures, Proceeding of IEEE,
Vol. 60, pp. 792 800, Jul. 1972.
D. Westerkamp, Adaptive Intra-Inter frame DPCMCoding for Transmission TV-Signals with 34 Mbps, IEEE
Zurich Seminar on digital Communication, pp. 39 45,
Mar. 1984.
M. H. Chan, Image coding Algorithms for Videoconferencing Applications, Ph.D. Thesis Imperial College
University of London, 1989.
J. O. Limb & J. A. Murphy, Measuring the Speed of
Moving Objects from television Signals, IEEE
Transaction on Communication, Vol. 23, pp. 474 478,
Apr. 1975.
C. Cafforio & F Rocca, Methods of Measuring Small
Displacements of Television Images, IEEE Transaction
on Information Theory, Vol. 22, pp.573 579, Sep. 1976.
A. N. Netravali & J. D. Robbins, Motion Compensated
television Coding ; Part 1, The Bell-System Technical
Journal, Vol. 58, pp. 631 670, Mar. 1979.
C. Cafforio & F Rocca, The Differential method for
Motion Estimation, Image Science Processing & Dynamic
Scene Analysis, Springer Verlag, New York, pp. 104
124, 1983.
J. D. Robbins & A. N. Netravali, Recursive motion
compensation : A Review, mage Science Processing &
Dynamic Scene Analysis, Springer Verlag, New York, pp.
75, 1983.
J. R. Jain & A. K. Jain, Displacement measurement & Its
Application in Interframe Image Coding, IEEE
Transaction on Communication, Vol. 29, pp. 1799 1808,
Dec. 1981.
T. Koga, K.Iinuma, A. Hirano, Y. Iijima & T. Ishiguro,
Motion-compensated Interframe Coding for Video
Conferencing, Proceeding National Telecommunication
Conference, New Orleans, LA., pp. G5.3.1 5.3.5, Nov.
1981.

ISSN: 2085-6350

[23]

[24]

[25]
[26]
[27]

[28]
[29]
[30]
[31]
[32]
[33]

[34]
[35]
[36]

[37]

[38]

Proceedings of CITEE, August 4, 2009

R. Srinivasan & K. R. Rao, Predictive Coding Based on
Efficient Motion Compensation, IEEE International
Conference on Communication, Amsterdam, pp. 521
526, May 1984.
A. Puri, H. M. Hang & D. L. Schilling, An Efficient
Block matching Algorithm for Motion Compensated
Coding, Proceeding IEEE ICASSP, pp. 25.4.1 25.4.4,
1987.
M. Ghanbari, The Cross Search Algorithm for Motion
Compensation, IEEE Transaction on Communication,
Vol. 38, pp. 950 953, Jul. 1990.
M. Gotze & G Ocylock, An Adaptive Interframe
Transform Coding System for Images, proceeding IEEE
ICASSP 82, pp. 448 451, 1982.
J. R. Jain & A. K. Jain, Displacement measurement & Its
Application in Interframe Image Coding, IEEE
Transaction on Communication, Vol. 29, pp. 1799 1808,
Dec. 1981.
J. A. Roese, W. K. Pratt & G. S. Robinson, Interframe
Cosine Transform Image Coding, IEEE Transaction on
Communication, Vol. 25, pp. 1329 1338, Nov. 1977.
J. A. Roese, Hybrid Transform predictive Image Coding
in Image Transmission Techniques, Academic Press, new
York, 1979.
CCITT H.261 Recommendations, Video Codec for
Audiovisual Services at p x 64 kbps, 1989.
G. Wallace, The JPEG Stil Picture Compression
Standard, Communication ACM, Vol. 34, pp. 30 44,
Apr. 1991.
D. LeGall, MPEG : A Video Compression Standard for
Multimedia Applications, Communication ACM, Vol. 34,
pp. 46 58, Apr. 1991.
S. G. Mallat, A Theory for Multiresolution Signal
Decomposition : the Wavelet Representation IEEE
Transaction on Pattern Analysis & Machine Intelligent,
Vol. 11, pp. 674 693, Jul. 1989.
S. G. Mallat, Multifrequency Channel Decomposition of
Image and Wavelet Models, IEEE Transaction on ASSP,
Vol. 37, pp. 2091 2110, Dec. 1989.
O Riol & M. Vetterli, Wavelet & Signal Processing,
IEEE Signal Processing Magazine, Vol. 8, pp. 14 38,
Oct. 1991.
Y. Q. Zhang & S. Zafar, Motion-Compensated Wavelet
Transform Coding for Color Video Compression, IEEE
Transaction on Circuit & Systems for Video Technology,
Vol. 2, pp. 285 296, Sep. 1992.
S. Zafar, Y. Q. Zhang & B. Jabbari, Multiscale Video
Representation Using Multiresolution Compensation &
Wavelet Decompostion, IEEE Journal Selected Area in
Communications, Vol. 11, pp. 24 34, Jan. 1993.
Project Team Tematics, Activity Report, INRIA, 2008.

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

CONTROL SYSTEM INTEGRATED STARTER - DC MOTOR COUPLE THREE PHASE INDUCTION MOTOR
FOR AUTOMOTIVE APPLICATIONS
Zulkarnain Lubis, Ahmed N. Abdalla
Faculty of Electrical and electronic Eng., University Malaysia Pahang, Kuantan 26300, Malaysia
[email protected]
[email protected]
Samsi bin MD said .Mortaza bin Mohamed
Faculty of Electrical and electronic Eng., TATI University College ,Kemaman 24000 Terengganu Malaysia.
Mortaza @ump.edu.my
[email protected]

Abstract-An electric vehicle control system controls motor

response based upon monitored vehicle characteristics to
provide consistent vehicle performance under a variety of
conditions for a given accelerator manipulation. With
emphasis on a cleaner environment and efficient operation,
vehicles today rely more and more heavily on electrical
power generation for success. With the oil price shocks of
the past few decades, as well as an increasing awareness of
the emissions of air pollutants and greenhouse gases from
cars and trucks, the interest to investigate alternative vehicle
propulsion systems has grown. This challenge of fuel
economy standards is promoting optimised and sometimes
novel vehicle power automotive architectures, which
combine the traditional internal combustion engine (ICE)
with various forms of electric drives. The different types of
the hybrid electric vehicles (HEV) are real competitors of
the classical ICE driven cars. The controller of induction
motor (IM) is designed based on input-output feedback
linearization technique. It allows greater electrical
generation capacity and the fuel economy and emissions
benefits of hybrid electric automotive propulsion. Finally, a
typical series hybrid electric vehicle is modelled and
investigated. Control system integrated starter dc motor
couple three phase induction motor for automotive
applications. Various tests, such as acceleration traversing
ramp, and fuel consumption and emission are performed on
the proposed model of 3 phase induction motor coupler dc
motor in electric hybrid vehicles drive.
Keywords: hybrid electrical vehicle, Induction motor, Dc
Machine.

I. INTRODUCTION
With the oil price shocks of the past few decades, as
well as an increasing awareness of the emissions of air
pollutants and greenhouse gases from cars and trucks, the
interest to investigate alternative vehicle propulsion systems
has grown. This challenge of fuel economy standards is
promoting optimised and sometimes novel vehicle
powertrain architectures, which combine the traditional
internal combustion engine (ICE) with various forms of
electric drives. The different types of the hybrid electric
vehicles (HEV) are real competitors of the classical ICE
driven cars.

In an all-electric vehicle (EV) there is no ICE, but all

other components exist including batteries with excessive
power. EVs and HEVs are studied by numerous authors in
the past, one comprehensive study is that of Chan [1]. First
full-scale hybrid vehicle work in Turkey is Doblo/Tofas
example realized at Marmara Research Center [2]. There
have been university theses and an industry project
constitutes the basics of this paper [3-7]. One of the main
contribution is that of Gokce [4], energy conservation and
energy balance method is adopted. The input-output
feedback linearization technique combined with an adaptive
backstopping observer in stator reference frame the
induction motor [5] using in series hybrid electric vehicle is
controlled.
This paper focus on a new HEV modelling to make a couple
two electric motor IM and DCM close loop sinusoidal
PWM inverter to control the speed of a three phase
induction motor. This compact inverter had its hardware
reduced to a minimum through the use of a programmable
integrated circuit (PIC) micro-controller (PIC16C73A). In
this sense a microcomputer interface was avoided. At the
end, a typical HEV is modelled and investigated. Simulation
results obtained show the IM and other components
performances for a typical city drive cycle.
2. Theoretical background
2.1 Management control system HEV
A hybrid electrical vehicle may consist of an internal
combustion engine (ICE), electric motor (EM), electric
generator (EG), power electronic circuits, advanced
electronic control units (ECU), a complex mechanical
transmission and a battery bank.
Fig.1 shows the structure of drive assembly of a hybrid
electric car. There are 3 electrical machines, generator and
starter (M/G), starter and the main motor (M), in the figure.
G/M is an integrated started and generator (ISG) which
connects with the internal combustion engine (ICE) using a
couple . The starter is a standby one. The M, which is
subject of this paper, is called main motor. It connects with
the wheels through the final gear. Main motor is a three
phase asynchronous Motor. The battery pack is a 288V,
10Ah NiH one. Fig.1

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

Proceedings of CITEE, August 4, 2009

The aerodynamic drag is due to the friction of the body

of vehicle moving through the air. The formula for this
component is as in the following .Dynamic modelling and
simulation of an induction motor with
fAD =

CD AV2 (2)

Fig.1 Management control system HEV

The hybrid electric car has 8 working modes: idle stop, ICE
drive motor drive, serial mode, parallel mode, serial &
parallel mode, ICE drive, battery charge and regenerative
braking. Fig.1 shows four of the modes. ICE stops running
when it is in the idle running state, and may be restarted in
less than 100ms by the M/G. The idle stop mode will reduce
fuel consumption and emissions in idle running state. The
ICE drive mode is the same as the traditional car and will
occur in most efficient working area of ICE. The motor
drive mode is the same as the battery electric car and will
occur at very low speed. In variasi mode which is shown in
Fig.1, the ICE drags the M/G to charge the battery, and the
main induction motor.
The first step in vehicle performance modelling is to
write an equation for the electric force . This is the force
transmitted to the ground through the drive wheels, and
propelling the vehicle forward. This force must overcome
the road load and accelerate the vehicle as shown in Fig.2

The gravity force due to the slope of the road can be

expressed by:
fgrade = Mg. sin (3)
Where is the grade angle.
In addition to the forces shown in Fig.3, another one is
needed to provide the linear acceleration of the vehicle
given by:
facc =M = M

(4)

The propulsion system must now overcome the road loads

and accelerate the vehicle by the tractive force, Ftot , as
follows :
Ftot = froll

fAD + fgarde + facc (5)

Whells and axels convert Ftot and the speed of vehicle

to torque and angular speed requirements for differential as
follow :
Twhell = Ftot rwheel , wheel = V/ rwheel (6)
Where Twhell , rwheel and wheel are the tractive torque, the
radius, and the angular velocity at the wheels, respectively.
The angular torque velocity and torque of the wheels are
converted to motor rpm and motor torque requirements
using the gears ratio at differential and gearbox as follows :
m = Gfd Ggb wheel , Tm = Twhell / Gfd Ggb (7)
Where Gfd and Ggb are respectively differential and gear
box gears ratios.

Fig.2Basicofforcesonavehicle

The rolling resistance is primarily due to the friction of

the vehicle tires on the road and can be written as:

3.PROPOSED METHOD
3.1.Controllers couple IM and DCM

froll = fr Mg , (1)
where M is the vehicle mass, f , is the rolling resistance
coefficient and g is gravity acceleration .

ISSN: 2085-6350

The coupling of these two components can be in parallel or

in series. In the parallel configuration, both the IM and the
DC electric motor contribute to the traction force that
moves the vehicle. Power is split between them according to
a control strategy, which is usually implemented by a

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

supervisory controller. Two different sub-controllers

independently control the IM and the DC motor. Both subcontrollers receive their commands from the supervisory
controller. Among these commands are the two torque
requests required from both sub-systems as shown in Fig.3.

Fig.5.Developedmodeloftransienttorqueandcurrentcontrolof
inductionmachine

Fig.3ControllersIMandDCMinatypicalHEVapplication

4. CONTROLLER SIMULATION MODEL

To study the performance of the developed Transient
torque and current model, a closed loop torque control of
the drive is simulated using Matlab/Simulink simulation
package. Fig.4 shows the simulation block diagram[9]. The
drive cycle gives the required vehicle speed then the torque
and speed requested from the electric motor. The current
drawn from IM power supply shows the battery
performance. The dynamic behaviour of the IM in the
DCM+IM drive cycle. Power assembly diagram of HEV
Normal Condition the ECE drive cycle. IM torque and
average torque, power assembly diagram of HEV in Hybrid
Electric. The block diagram of the simulink model is shown
in Fig.4.

5. Proposed control system

5.1 Simulation model of induction motor couple dc
motor
Simulation of the IM couple DC motor drive system is
performed
in
Matlab/Simulink
and
SimPower
environments. Voltage sags of types A-G, produced by
software, are applied to the test system. Schematic diagram
of the simulated system is shown in Fig. 6. The point here is
to show how easy it is to take the SIMULINK blocks from
the Library and turn them into a simulation and then into a
real-time implementation

Fig.6 Three phase induction motor couple dc motor

Fig.4.Simulationblockdiagramforstabilitycontrol

changing conditions. Careful choice of the method of

combining the controllers may result in a highly adequate
yet non-oscillatory response. To take advantage of the rapid
response of the PI-type controller, one needs to keep the
system responding under the PI controller for a majority of
the time and use the electric controller only when the
system behaviour is oscillatory or tends to overshoots

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

Proceedings of CITEE, August 4, 2009

10

5.2 Experiment setup

This paper describes the development of a experiment
library of control and modelling algorithms for the various
types of induction motor drives considered for Hybrid
Electric Vehicles (HEVs). The wiring diagram control of
three phase induction motor coupler dc motor in Fig.7.b and
the set up experiment LAB in Fig.7.a

a. Experiment set up

b. Wiring diagram control of

three phase induction motor
coupler dc motor

Fig. 7 Experimental wiring diagram control

6. SIMULATION AND. EXPERIMENTAL RESULTS

6.1. THE SIMULATION RESULTS
To demonstrate the proposed hybrid Electric control
scheme success, it has been tested by simulation, in order to
evaluate the performances under a variety of operating
conditions. The numerical values for the tested induction
motor are summarized in Table I.

electric controller reduces both the overshoot and extent of

oscillations under the same operating conditions. To
demonstrate the robustness of the proposed controller a
different type of trajectory was considered in this test. High
tracking accuracy is observed at all speed. One can see from
these figures that the results using electric controller, were
very successful.
To illustrate the effectiveness of the switching strategy
further, the hybrid controller was applied to control the
motor under variable load torque. It is observed from Fig.8,
that the hybrid controller closely tracks the motor speeds,
even under changing conditions. Rejection of an external
disturbances is also achieved. Compared with the motor
speed response with variable load, it can be seen that the
undesirable oscillatory response is clearly evident.
All test results show that the proposed hybrid electric
control strategy is very effective in tracking the selected
tracks at all time, while the system transients are effectively
reduced. The results presented in Fig. 8 show that the
proposed control system works correctly. The plots of these
figures show the performance as the proposed scheme of
hybrid-electric controller for variety of step changes in the
desired set point. It can be observed that, the application of
external force of 1.0 (N.m) to induction motor, the control
and set-point following are satisfactory. In order to examine
the robustness of the proposed control scheme, we assume
that the parameters of rotor resistance Rr and load inertia J
have been perturbed from their nominal values.

Table 1 Rating of tested Induction motor

The controller algorithm is housed inside the personal

computer with Pentium-4 microprocessor and all numerical
values of the simulation model are obtained either by
measurements. The software environment used of these
simulation experiments is Matlab-software with Simulink
Toolboxes. For all simulations performed in this paper, the
best gain, found experimentally to be kp =0.56 and ki =
10.04.
After designing the best stand alone PI and electric
controllers, all effectiveness of combining the two
controllers to produce a hybrid design is demonstrated.
Simulation results are given for motor sped tracking with
the desired speed changing from the level to another
(square-wave reference track with amplitude 150 rad/s).
Figs. 8 show the speed trajectory when the desired
speed changes from one value to another, using the PI
controller and the electric controller, respectively. The
measured speed is superimposed on the specified desired
speed in order to compare tracking accuracy. Clearly, the
ISSN: 2085-6350

Fig.8 Speed control system of Induction motor couple dc motor using PI

Controller and electric controller Speed step response under load and
rotor resistance changes.

The parameters of stator resistance, inductances and viscous

friction f maintain their nominal values. It is evident that the
speed response of the proposed control scheme is not
significantly affected by these variations. One can see from
these all figures the results were very successful and the
obtained results confirm the validity of the proposed control
scheme. These figures reveal that the proposed controller
based on the hybrid electric scheme was superior to the
conventional controllers, such as the Proportional Integral
PI-type. The controller envisaged is capable of maintaining

Conference on Information Technology and Electrical Engineering (CITEE)

11

Proceedings of CITEE, August 4, 2009

a high tracking accuracy even in the presence of sudden

disturbances such as load of electric transients.

good speed response regardless of parameter variation or

external force.

6.2. THE EXPERIMENTAL RESULTS

Reference

Results is explained by the pact that, in experimental

test were observed a strong influence of motor inductance in
coupler to dc motor, more precisely, in the power system,
however in the simulation such influence was not
considered, and also non-linearity and additional losses.

[1] C. C. CHAN, The State of Art of Electric and Hybrid

Vehicles,

Proceedings of IEEE, vol. 90, no. 2, Feb. 2002.

[2] C. GOKCE, Modeling and Simulation of a Series Parallel Hybrid
Electrical Vehicle, Master Thesis, Institute of Science and
Technology, Istanbul Technical University, Istanbul 2005.
[3] T. Markel, A. Brooker, T. Hendricks, V. Johnson, K. Kelly, B. Kramer,
M. O' Keefe, S. Sprik, K. Wipke: ADVISOR: A Systems Analysis
Tool for Advanced Vehicle Modelling, ELSEVIER Journal of Power
Sources 110, (2002), pp. 255-266.
[4] H.F. Abdul Wahab and H. Sanusi Simulink Model of Direct Torque
Control of Induction Machine American Journal of Applied Sciences
5 (8): 1083-1090, 2008 ISSN 1546-9239. 2008 Science Publications.

a.

[5] Stephen Yurkovich Nonlinear Torque Control of the Induction Motor

in Hybrid Electric Vehicle Applications,American Control
Conference ,June 8-10, 2005. Portland, OR, USA
[6] RomeOrtega, Nildta Uarabanov, Gerarda Escabar Valderrama. Direct
torque control of induction motors:stability analysis and performance
improvement, IEEE Trans. on Auto. Control, Vol. 46 (8), 2001, pp.
1209-1221.
[7] R. Yazdanpanah, A. Farrokh Payam: Direct Torque Control of An
Induction Motor Drive Based on Input-Output Feedback Linearization
Using Adaptive Backstepping Flux Observer, Proc. 2006 AIESP
Conf., Madeira, Portugal.

b.
Fig. 9 Speed & torque control system of a. Only Induction motor and
Induction motor couple dc motor

[8] S. Sadeghi, J. Milimonfared, M. Mirsalim, M. Jalalifar: Dynamic

Modeling and Simulation of a Switched Reluctance Motor in Electric
Vehicle, in Proc., 2006 ICIEA Conf.

7. Conclusions
In this paper, the couple of the two electric motors with
input-output state feedback controller combined with
adaptive back stepping observer and batteries of a typical
series HEV is investigated and simulated by
Matlab/Simulink, has been presented and the performance
and ability of control strategy is investigated. The proposed
control system was analysed and implemented and its
effectiveness in tracking application was verified. From the
above results it is clear that the controller despite of its
simple structure has all of the futures of a high precision
speed controller for operating in the whole of speed range
and for any loading and environmental conditions and had a

[9] C. GOKCE, O. USTUN, M. YILMAZ, R. N. TUNCAY, Modelling

and Simulation of Series Parallel Hybrid Electrical Vehicle,
ELECO05, Int. Conference on Electrical and Computer Engineering,
Bursa, 2005.
[10] C. Won, S. Kim and B.K. Bose, Robust position control of induction
motor using fuzzy logic control. In Con. Rec. IEEE-IAS Annu.
Meeting, Conf. Rec, pp.472-451, Oct. 1992.
[11] L. Zadeh. Outline of a new approach to the analysis of complex
systems and decision processes. IEEE Trans. Syst., Man, Cyben.,
Vol.3.pp.28-44, Jan. 1973.
[12] Ioan Adrian Viorel, Lornd SZAB Integrated Starter-Generators
For Automotive Applications, Siemens , Acta Electrotehnica,
Transportation Systems,Vol.45, No.3, 2004, Erlangen, Germany.

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

12

Proceedings of CITEE, August 4, 2009

Transient Stability of SMIB: a Case Study

Adelhard Beni Rehiara
Engineering Department, University of Papua (UNIPA)
Jl. Gunung Salju Manokwari, 98314, Indonesia
email : [email protected]
Abstract Single machine infinity bus (SMIB) is a simple
way to examine a complex electrical power system. In
investigating the transient stability of a SMIB system, equal
area criterion (EAC) method can be used to get critical
clearing angle cr and critical clearing time tcr. In each case of
a SMIB, critical clearing time tcr cannot directly be determined
using equal area criterion method. This paper will introduce
Runge-Kutta method utilized to modify the critical clearing
time tcr found with EAC method and to know the best time to
clear a fault. In this case, the critical clearing time tcr of EAC
method is almost same for every fault and it is faster than the
critical clearing time tcr of Runge-Kutta method.
Keywords Transient stability, single machine infinity bus,
equal area criterion, Runge-Kutta, step by step.

will be stable if the acceleration area A1 is smaller or equal

to the deceleration area A2. According to the assumption,
critical clearing angle cr and maximum power angle max
can be found as follow [10],[11],[13].
Pm(max ) P maxII cos + P maxIII cosmax

P maxIII P maxII

cr = cos1

max = sin 1 Pm P max

INTRODUCTION

Electrical power systems consist of generation,

transmission and distribution system and/or also load as the
user of the electrical power. The other components that can
probably be connected to the systems are transformer, circuit
breaker, relay protection, prime mover, etc. All of the power
system components are used to maintain the quality,
continuity, stability and reliability of the systems.

(2)

The critical clearing time also can be established in EAC

method as shown in equation 3.
tcr =

I.

III

(1)

2H(cr )
f Pm

(3)

Where H is the stored kinetic energy at synchronous

speed per MVA base, f is the frequency in hertz, Pm is the
shaft power input less rotational losses in pu, t is the time in
second and Pmax is the maximum electrical power in pu
while the subscript I, II and III indicate the condition of
steady state (pre), during and post fault.

SMIB (single machine infinite bus) is a simple electrical

power system that has a generator connected to infinite bus
as load [11]. To make it simple to be analyzed, an
interconnection electrical power system can be separated
into some SMIBs.
Stability is an important constraint in power system
operation [2]. The major problem in every electrical power
system is how stability of the system when a fault happens
for example short circuit, broken line, disconnected load, etc.
Equal area criterion (EAC) is a classic method in
transient stability that is applicable to all two machines
systems (SMIB). The method provides an easier way to
determine the critical clearing angle cr but the method does
not able to find directly the critical clearing time tcr [4],[13].
Another mathematical calculation, called Runge-Kutta
methods, will be used to solve the limitation of the EAC
methods. This paper will combine both EAC and RungeKutta methods to get the best critical clearing time tcr.
II.

TRANSIENT STABILITY

A. Equal Area Criterion

The Equal area criterion method is based on an
assumption that acceleration area A1 should be same with
deceleration area A2 (figure 1). The areas are separated
between normal power angle and maximum power angle
max equally by critical clearing angle cr. A power system

Figure 1. Power curve

B. Runge-Kutta Method
The Runge-Kutta method is a mathematical method to
solve the differential equation in every step of time (step by
step). The step by step calculations will be used to find
critical clearing time tcr according to critical clearing angle
defined with the EAC method. The power angle is
calculated for several times to ensure that the angle will be
oscillated indicating that the system is stable.
Curve versus t of a machine should be made to
investigate the swing curve of a machine. By determining
the swing curve in various clearing time, the maximum time
permitted to clear the fault can be determined.

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

Proceedings of CITEE, August 4, 2009

13

Some numeric methods, which are often used to solve

the differential equation of step by step calculation, are the
methods of Euler, Heun, Runge-Kutta, etc. In this case, it
will be focused on the method of Runge-Kutta.
Fourth order of Runge-Kutta method can be utilized to
analyze the swing equation. The equation can be rewritten as
[3],[6]:
d
= s
dt

(4)

Where Pe is the electrical power in pu, Pa is the

accelerating power in pu and is the angular displacement
of the rotor in rad. By substituting the swing equation to the
method of Runge-Kutta, four estimations can be obtained
[6].
First estimation:

d
dt

d
dt

t = ( t ) s t
(t )

t =
(t )

f
Pm Pe(( t1)) t
H

(6)

f
Pm Pe(( t2)) t
H

l
= ( t ) + 2i s t
2

Fourth estimation:

k 4i =

l 4i

RESULTS

The single line diagram of an investigated power system

is shown in the figure 2. The system consists of a generator
attached to two of step up transformers and step down
transformers connected to infinite bus, two transmission
lines and a load.
j0.15

j0.28

j0.15

j0.25

j0.17

V=1.0 pu
j0.15

j0.28

j0.15

A. EAC Calculations
The EAC calculations will be separated into pre, during
and post fault condition.

(7)
(8)

Third estimation:

l 3i =

(15)

c+1 c

Figure 2. Single line diagram

l
k 2i = ( t ) + 1i s t
2

k 3i

( c+1 + c ) c t

(5)

Second estimation:

l 2i =

Where tc and tc+1 are the time for clearing and after
clearing fault while c and c+1 are the angle at clearing and
after clearing fault.

E=1.2 pu

k 1i =

III.

d
(Pm Pe ) = Pa
=
dt
2H
2H

l1i =

t cr = t c +

(9)

(10)

{(

(11)

+ l 3 i s t

f
Pm Pe(( t4)) t
=
H

(12)

Where i = 1, 2, 3, ... , n and in the end of the period, the

power angle and the synchronous speed will be changed
using both equations below.

i(t + t ) = (t ) + 1 6 (k li + 2k 2i + 2k 3i + k 4i )

(t + t ) = (t ) + 1 6 (lli + 2l2i + 2l3i + l4i )

To get the electrical power Pe, a transfer reactance Xt,

which is the reactance from the generator to the load, should
be known. From the figure 2, the transfer reactance Xt can
be calculated as follows:
Xt = j 0.25 +

f
Pm Pe(( t3)) t
H

(t )

1) Pre fault

(13)
(14)

The calculation will be continued for t=t+t until the

required duration of time. Critical clearing time tcr can be
estimated using a linear interpolation method, which is
formulated as [1]:

j (0.15 + 0.28 + 0.15) j (0.15 + 0.28 + 0.15)

+ j 0.17
j (0.15 + 0.28 + 0.15) + j (0.15 + 0.28 + 0.15)

= j 0.71

V E

1 .2
sin =
sin
Xt
0.71
= 1.69014 sin

Pe =

With shaft power input (Pm) is 1.0 pu, the steady state
power angle is:
= sin 1 (1 / 1.69014)
= 36.2754 = 0.6331rad

2) During fault
Because the reactance of step up and step down
transformers and also transmission lines are same, the
calculation of during fault can only be done for line 1 or 2.
The reactance system after -Y conversion is shown in
figure 3. The transfer reactance of each fault is found by
converting it back with Y- conversion. The transfer
reactance and maximum power of each fault are shown in
table I.

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

14

Proceedings of CITEE, August 4, 2009

j0.25

j0.0556

j0.2150

j0.17

j0.0750

V=1.0 pu

E=1.2 pu
(a) Front line fault

j0.25

j0.0145

j0.145

j0.17

j0.0725

V=1.0 pu

E=1.2 pu
(b) Midle line fault
j0.25

j0.0556

j0.0750

j0.17

B. Step by Step Calculations

The system is assumed to work with frequency 50 Hz
and H=4MJ/MVA and the iteration interval is 0.05s. For the
first time the fault happens, the acceleration power Pa is not
in synchronous. So the acceleration power Pa is the average
of Pa pre and Pa during fault.
1) Fault in a front line
The step by step calculation was done in Matlab with
several clearing times tc and it was presented in following
figures.

j0.0250

180

V=1.0 pu

E=1.2 pu

tc = 0.05s

(c) End line fault

3) Post fault
When the fault is cleared, the system is operated by using
a line. So the transfer reactance Xt and electric power Pe will
be changed and those can be calculated as:
Xt

power angle (deg)

Figure 3. -Y conversion

tc = 0.10s
tc = 0.15s

140

tc = 0.30s

120
100
80
60
40

= j0.25 + j0.15 + j0.28 + j0.15 + j0.17

= j1.0
= 1.2 sin

Pe

160

20
0

The maximum power angle max and the critical clearing

angle are:

( )

2.5

tc = 0.05s

1(2.1565 0.6331) 0.4054cos 0.6331+1.2 cos 2.1565

cr = cos

1.2 0.4054

= 0.835191rad = 47.8529

8(47.8529 36.2764)
50
= 0.101s

tcr =

synchronous speed (rad)

In EAC method, the critical clearing time tcr is offered

with equation 3 and the result for fault in front line is:

tc = 0.10s
tc = 0.15s

350

tc = 0.30s

340

330

320

310

300

Power angle of normal operation and maximum power

Pmax of pre and post fault are same wherever the fault
happens. Result of calculations for the EAC method is given
in table I.
VARIABLES DEFINED WITH EAC METHOD
Fault Location
Front

Middle

End

Xt (pu)

j2.960

j2.4262

j2.759

Pmax (pu)

0.4054

0.4946

0.4349

cr (deg)

47.8529

49.1602

48.2543

0.101

0.107

0.103

tcr (s)

1.5
time (s)

360

= 2.1565 rad

Variables

Figure 4. Curve of for fault in front line

max = sin 1 11.2

TABLE I.

0.5

0.5

1.5
time (s)

2.5

Figure 5. Curve of for fault in front line

The results of the calculations are angle and

synchronous speed as shown in the figure 4 and 5. Base on
the figures, the best clearing time tc is about 0.1s because
clearing time a step toward (0.15s) will make the system
unstable. So the critical clearing time tcr is probably in
between 0.1s and 0.15s.
2) Fault in a middle line
The calculation results of the middle line fault are figured
on figure 5 and 6. The calculations were done with PeI=1.69
sin , PeII=0.4946 sin and PeIII=1.2 sin . With clearing
time tc = 0.15s, the swing curve is wide but the system is still
stable.

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

Proceedings of CITEE, August 4, 2009

15

Based on the next figure 6 and 7, the best time to clear

the fault in the middle line is in 0.15s. The system will be
unstable for the next step.

360
tc = 0.05s
tc = 0.10s
tc = 0.15s

350

synchronous speed (rad)

180

power angle (deg)

tc = 0.10s

160

tc = 0.15s
tc = 0.20s

140

tc = 0.30s

120
100
80

tc = 0.30s

340

330

320

310

60
300
40

20
0

0.5

1.5
time (s)

2.5

1.5
time (s)

2.5

tc = 0.10s
tc = 0.15s
tc = 0.20s

350

4) Critical Clearing Time

To find the critical clearing time tcr, every fault will be
catch for a few milliseconds and find the critical clearing
time tcr by checking the critical clearing angle cr in each step
of time duration.

360

synchronous speed (rad)

Figure 9. Curve of for fault in end line

Figure 6. Curve of for fault in middle line

The results of the step by step calculations are shown in

table II. The results are only provided the data for a half of
first second calculations. The shaded areas inside the table
indicate that the critical clearing time tcr is probably in
between the step.

tc = 0.30s

340

330

For fault in front line, the critical clearing angle tcr is

about 47.8529 degree and it is in between step 0.1s and 0.15s
(see table II). The critical clearing angle tcr can be estimated
by using the equation 15 with data provided in table II and it
can be calculated s follows.

320

310

300

0.5

0.5

1.5
time (s)

2.5

tcr = 0.10 +

Figure 7. Curve of for fault in middle line

(48.0217 + 41.3577) 41.3577

x 0.05
48.0217 41.3577

= 0.125 s

3) Fault in an end line

Figure 8 and 9 show that the fault in the end line should
be cleared at 0.1s because the system will be unstable for the
next step (0.15s).

With the same method, the critical clearing time tcr for
fault in the middle and the end line are found about 0.175s
and 0.125s.

180

power angle (deg)

tc = 0.05s

160

tc = 0.10s
tc = 0.15s

140

tc = 0.30s

TABLE II.

120
100
80
60
40
20
0

0.5

1.5
time (s)

Figure 8. Curve of for fault in end line

2.5

00+
0avg
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50

RESULTS OF THE STEP BY STEP CALCULATIONS

Front line fault

(deg)
36.2754
36.2754
36.2754
37.2685
41.5061
49.8488
62.0701
77.9062
97.1527
119.7880
146.1051
176.8129
213.0398

(deg)
0.9931
4.2376
8.3428
12.2213
15.8361
19.2465
22.6354
26.3170
30.7078
36.2269
43.0643

Conference on Information Technology and Electrical Engineering (CITEE)

Middle line fault

(deg)
36.2754
36.2754
36.2754
37.2685
41.2027
48.9193
60.1594
74.6150
92.0274
112.3098
135.6790
162.7707
194.6991

(deg)
0.9931
3.9342
7.7167
11.2400
14.4556
17.4124
20.2825
23.3691
27.0917
31.9284
38.2611

End line fault

(deg)
36.2754
36.2754
36.2754
37.2685
41.4057
49.5408
61.4358
76.8116
95.4456
117.2956
142.6328
172.1495
206.9836

(deg)
0.9931
4.1372
8.1351
11.8950
15.3758
18.6340
21.8501
25.3371
29.5167
34.8341
41.5487

ISSN: 2085-6350

16

Proceedings of CITEE, August 4, 2009

IV.

CONCLUSSIONS

Critical clearing time tcr found with EAC method is

almost same for the investigated system wherever the fault
happens. For each fault in the system, the critical clearing
angle cr is different because the electrical power Pe for
every fault is different. The difference also makes an
influence for the critical clearing time tcr.
Using Runge-Kutta method, critical clearing angle cr
can not be used to define the critical clearing time tcr because
the calculation using step by step is never exactly match with
the critical clearing angle cr. In this case, the critical
clearing time tcr can be found with a linear interpolation
method.
The result shows that the critical clearing time tcr defined
with EAC method is faster than with Runge-Kutta method.
Some differences appear in the result but both EAC and
Runge-Kutta methods prove that the critical clearing time tcr
is longer for every fault in the middle line.

[3]
[4]
[5]
[6]
[7]

[8]
[9]

[10]
[11]
[12]

REFERENCES
[1]
[2]

Achmad Basuki, Metode Numerik dan Algoritma Komputansi,

Andi Offset, Yogyakarta, 2004.
Deqiang Gan, Robert J. Thomas, Ray D. Zimmerman, A Transient
Stability Constrained Optimal Power Flow, Bulk Power System

[13]

Dynamics and Control IV Restructuring, Santorini, Greece, August

24-28, 1998.
E. W. Kimbark, Power System Stability, vol. 2, John Wiley and
Sons, Inc., New York, 1995.
Elhawary, Mohamed E., Electrical Power Systems; Design and
Analysis, Reston Publishing, Reston VA, 1995.
Fitzgerald, A.E.Kingsley, C. Umans, Mesin-Mesin Listrik,
Erlangga, Jakarta, 1997.
Glenn W. Stagg, Ahmed H. Abiad, Computer Method in Power
System Analysis, McGraw-Hill, 8th. edition, Kogakusha, 1984.
Khan, A PC Based Software Package for The Equal area Criterion
of Power System Transient Stability, IEEE Trans. Power System,
Vol. 13, No. 1, Feb. 1998.
M.A.Pai, Computer Techniques in Power System Analysis, Tata
McGraw-Hill, New Delhi, 2006.
Mohammad Reza, A Survey on the Transient Stability of Power
Systems with Converter Connected Distributed Generation, Jurnal
Teknik Elektro Vol. 7, No. 1, March 2007.
Nagrath, D.P. Kothari, Modern Power System Analysis, second
edition, Tata McGraw-Hill, New Delhi, 2001.
Prabha Kundur, Power system Analysis, Tata McGraw-Hill, New
Delhi, 1994.
Shengli Cheng, Mohindar S. Sachdev5, Out-of-Step Protection
Using the Equal Area Criterion, IEEE CCECE/CCGEI, Saskatoon,
May 2000.
William D. Stevenson Jr, Analisis Sistem Tenaga Listrik, Erlangga,
Jakarta, 1996.

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

17

Proceedings of CITEE, August 4, 2009

LOOK-UP TABLE OF FUZZY RULE SURAM

WITH AVR ATmega128
Zakarias Situmorang
Faculty of Computer Science, Catholic University of Saint Thomas Medan
Jl. Setiabudi No. 479-F Tanjungsari Medan 20132
Email : [email protected]

ABSTRACT
Implemented of fuzzy rule must used a look-up table as
defuzzification analysis. Look-up table is the actuator plant
to doing the value of fuzzification. Rule suram based of
fuzzy logic with variables of weather is temperature
ambient and conditions of air is humidity ambient, it
implemented for wood drying process. The membership
function of variable of state represented in error value and
change error with typical map of triangle and map of
trapezium. Result of analysis to reach 8 fuzzy rule in 150
conditions to control the output system can be constructed
in a number of way of weather and conditions of air. It used
to minimum of the consumption of electric energy by
heater. One cycle of schedule drying is
a serial of
condition of chamber to process as use as a wood species.
Design in control used a AVR Atmega-128 as has a
memory very big to apply a source code of a wood process
schedule of drying.
keywords : look-up table, defuzzification, fuzzy controller, a
wood schedule of drying, AVR Atmega-128

Introduction
The wood drying process used the schedule of
drying dependent for moisture of content the wood, that
condition of kiln in temperature and humidity of chamber.
The controller used to control the actuators are heater,
sprayer and damper, whenever the process used doing the
optimal from time and energy and stability in wood
schedule of drying. Main source of energy is solar energy
from collector and alternative source energy by heater.
Number of solar energy based of intensity of solar and
alternative energy by heater is consumption of electric.
The maximized use of solar energy in wood drying
process is goal of control system. Its depended by a
number solar energy and its change and temperature of
ambient. Responsibility of the change the solar energy in
variable of temperature ambient and humidity ambient is the
especial of goal the control system. The process of control
is to hope maximized the use of solar energy and the
minimized of consumption energy of electric.
Process of wood drying is depended for schedule
drying, which used to track of set point for temperature and
humidity drying. The conditions of temperature and
humidity drying in schedule are different for the each steps
of the schedule drying of wood. Variable control the Wood
drying of process kiln are temperature and humidity of air in

chamber where dependent for moisture of content the wood.

Its need to control of actuator system for heater, sprayer
and damper, whenever the process used doing the optimal
from time and energy and the actuator doing in the
conditions data riel.
The Fuzzy rule suram implemented for schedule
drying of Albasia Albizia wood and
modification of
membership function in range [0.5, 1].
Wood Drying Process
Control variable of a solar energy wood drying
process kiln is temperature and humidity to adapt variable
of a drying schedule. Dimension of wood drying kiln has
designed and built several type dry kiln for use lumber of
housing structure. The Schedule drying is a cycle of drying
and have the some level of process. A Level process doing
at temperature and humidity variable are constant at set
point any time. By the way need an actuator control system
(heater, sprayer and damper) then doing at effective the time
and efficiency of energy.
Air drying involves the open piling of freshly
processed timber in stacks out of doors or in open sheds so
that the wood surfaces are exposed to the surrounding
atmosphere. During air drying there is little control over the
factors that influence drying and hence the rate of drying is
very much controlled by local atmospheric conditions. One
guiding principle during air drying of timber is to ensure
adequate air circulation through the stack. In practice this is
achieved by separating each board within the stack using
sticks or stickers
High temperature drying, ie., drying above the
boiling point of water, is the dominant method for the
drying of radiata pine (Industry Standard 100-1992). The
method which is used to dry wood in chamber with some
level process is described below;
a. Pre-Heating Process
This process describe to stability of temperature
distributed in wood stacks is equally. . The rate of drying
can be controlled to some extent by altering the size of
the sticks. Drying can be slowed by using thinner sticks
or increased by using wider sticks. During air drying, air
circulating within the stack absorbs moisture, is cooled
and in this process drops to the bottom of the stack. If a
space or chimney is constructed within the centre of the
stack then this cooler air can exit the stack and fresh air
will be drawn in at the top of the stack. If this process
occurs correctly then air movement and drying

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

Proceedings of CITEE, August 4, 2009

18

concentration of diffusing molecules at the board surface,

simply by increasing the temperature within the kiln. The
loss of water from the board surface to the air is by
evaporation. Evaporation is temperature dependent and
relies on a relative humidity of the air surrounding the
timber. If hot dry air is used for drying then the surface
layers of the drying timber may dry rapidly and develop
case hardening or checks. To avoid excessive evaporation
from board surfaces the air in the kiln must be kept
relatively humid. Kiln drying therefore involves careful
control of both air temperature and humidity within the kiln.
Air circulation as in air drying also plays an important role
in the control of humidity as it ensures that the air in contact
with the timber is renewed frequently thereby preventing
the development of a stagnant layer of excessively humid
air from occurring around the timber.
A drying schedule is a set of directions for the
operation of a kiln during the drying period and is usually
presented in the form of tables showing the temperatures
and humidity to be used at various stages of the process.
These schedules vary with species, size and grade of timber
being dried; they are also influenced by local production
and selling practices, by the degree of care in kiln operation
and in kiln characteristics. Schedule drying for Albasia
albazia wood is shown in figure 1

throughout the stack should be continuous and uniform.

Humidity drying make to between 80% - 90% and heater
and sprayer are doing for 2 hours per 1 cm tick of wood.
b. Drying Process
Immediately after plasticisation the timber is dried until
its moisture content is 6% on average and 10% maximum
with the kiln conditions as follows. Maximum dry bulb
temperature between 116-160 degrees centigrade
Minimum wet bulb temperature
40-50 degrees
centigrade. Minimum average air velocity within the
stack, 4.0 m/sec, 5.0 may be the optimum
c. Equalizing and conditioning Process
Immediately following drying, the timber is cooled to a
surface temperature of between 90-100 degrees
centigrade and then steamed under saturated conditions
for four hours at a temperature of not less than 95 degrees
centigrade.
Movement of water within wood can occur through a
number of processes including capillarity, vapour diffusion,
bound water diffusion, combined vapour diffusion and
bound water diffusion and at temperatures in excess of 100
degrees centigrade mass flow. Diffusion processes are
increased with increasing temperature and therefore it is
possible to increase the rate of water movement within the
board to the board surface, assuming there is a lower

Schedule Drying Albasia Albazia Wood

80
Humidity Drying (%)

Wood MC (%)
Temperature Drying (oC)

90
70
60
50
40
30
20
10
0
1
292 583 874 1165 1456 1747 2038 2329 2620 2911 3202 3493 3784 4075 4366 4657 4948 5239 5530
Moisture Content
Temperatur Drying
Humiditi Drying
Time (menit)

Figure. 1. A Schedule Drying for Albasia albazia wood

Solar Energy in a Prototyping of Wood Drying Kiln
Mearsurement of solar energy by Piranometer type
MS-801 Chino, it has maximum voltage +50 mVDC, which

to have a data from Agency of meteorology and geophysics

Yogyakarta Indonesia , its shown in figure 2. with detail of
a prototyping of wood drying.

9.Piranometer

6.sprayer
5.Damper in
2.Door

3.heater
7.colector

110 cm

80 cm

4.fan
8.glases
50 cm

1.wall
5.Damper out

60 cm
15 cm

50 cm

10 cm

Figure. 2. Detail of design the prototyping of wood drying kiln

ISSN: 2085-6350

Conference on Information Technology and Electrical Engineering (CITEE)

19

Proceedings of CITEE, August 4, 2009

Most important part of a dehumidifying kiln is a

heat pump. Air in the kiln passes through the load of timber
and absorbs moisture from the wood. Part of that air
circulates through the heat pump where the moisture is
condensed and drained out of the chamber. Dried, reheated
air comes back to kiln chamber. Energy consumption is

minimal, since there is almost no inner/outer air exchange.

Air circulation in these kilns is smaller than in conventional
kilns while the electrical power of flow fans and heat pump
equals approximately the power of flow fans in
conventional kilns.

A number of variable solar radiation can be convert to variable temperature ambient and humidity ambient shown in table 1.
Table. 1. Convert solar radiation to variable temp. ambient and humidity ambient for Albasia albizia wood.

1
1
2

Set
Point
Td0
(oC)
2
45
50

55

60

65
(F)

No.

Set
Point
Hd0
(%)
3
60
55
50
45
40
35
60

MC (%)
4
25.5 - 50
10.4 25.5
7.8 10.4
7.3 7.8
6.7 7.3
6.4 6.7
6.3 - 6.4

Solar
Radiation
Io
(Watt/m2)
6
561.0
662.0

Temp. Ambient Ta(oC)

Hum. Ambient Ha (oC)

Min

Rate

Max

Min

Rate

Max

7
29.1
29.7

8
30.1
30.7

9
31.1
31.7

10
52
46

11
60
54

12
68
62

763.0

30.5

31.5

32.5

40

48

56

617.5

29.4

30.4

31.4

50

58

66

535.0

28.9

29.9

30.9

54

62

70

Note :
i. For MC 25 %,( Td0 = 45 s/d 55), The Solar Radiation I 0 = 460 + 20.2(Td o 40) (watt/m2)
ii.

For MC < 25 %,( Td0 = 60 s/d 70), The Solar Radiation I 0 = 700 16.5(Td o 55) (watt/m2)

iii. Control priode for total used the energi from solar , Cp = 0.1 + 0.0003( I 0 400) (0C/menit)
F = Fakultatif

Fuzzification of membership function in range [0.5, 1],

and for weather (temperature of ambient Ta and change
temperature ambient CTa) for set point temperature drying
Td0 = 450C and Humidity drying Rd0 = 60%) describe by
figure 3. with variable M : over-cloudy ; B : cloudy; CB :
bright-cloud ; C: clear; SC : clearest and for change of
temperature ambient used -H = - High, -M = -Medium, -S =
- Small, Z = Zero, S = Small, M = Medium, H = High.

Design Membership Functions

Design of membership function for variables in fuzzy
logic implemented to mapp of typical of triangle and
trapezium as of a drying schedule as shown in figure 1. and
table 1 Its needed to implemented control system of wood
drying kiln.

1
1

CB

28.1

29.1

30.1

31.1

SC

-H B -M

-S

+S

+M

0.2

0.4

+H

0.5

0.5

Figure. 3. Membership function of temperature ambient and

change temperature ambient for Td0 = 450C and
Rd0 = 60%

Computation process of variable change of temperature

ambient are given eq.1. with n 0 s/d

CTa(0C)
- 0.6 - 0.4 - 0.2

32.1 Ta(0C)

0.6

CTa [(n+1)T] = Ta[(n+1)T] Ta[nT]

(1)

Representative of weather is variable temperature

ambient and this change used to maximize and to hoist
responsibility of membership Function, with appropriate
rule in a table look-up scheme, shown table 2.

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

Proceedings of CITEE, August 4, 2009

20

Table 2. Look-up table for membership Function for temperature of ambient

Cloudy
Bright-Cloud
Clear
Ta
Over-cloudy (M)
(B)
(CB)
(C)
CTa
Over-cloudy
Over-cloudy
Cloudy
Bright-Cloud
-H
Over-cloudy
Cloudy
Bright-Cloud
Clear
-M
Over-cloudy
Cloudy
Bright-Cloud
Clear
-S
Over-cloudy
Cloudy
Bright-Cloud
Clear
Z
Over-cloudy
Cloudy
Bright-Cloud
Clear
+S
Over-cloudy
Cloudy
Bright-Cloud
Clear
+M
Cloudy
Bright-Cloud
Clear
Clearest
+H

Fuzzyfikasi of membership function in range [0.5,

1] of conditions of air (humidity of ambient Ha and change
of humidity ambient CHa) for set point temperature drying
(Td0 = 450C) and humidity drying (Rd0 = 60%) describe by

Clearest
(SC)
Clear
Clearest
Clearest
Clearest
Clearest
Clearest
Clearest

figure 4. with variable : P : Hot; AP : Rather-Hot; H :

Swarm; S: Fresh; D : Cold, and for change of humidity
ambient used -H = - High, -M = -Medium, -S = - Small, Z =
Zero, S = Small, M = Medium, H = High.

B
1

AP

52

56

60

64

68

-H B -M

0.5

-S

+S

+M

+H

-1

0.5

CHa(%)
-3

Ha(%)

-2

Figure. 4. Membership Function of humidity ambient and change humidity for Td0 = 450C and Rd0 = 60%

Computation process of variable change of humidity

ambient are given eq.2. with n 0 to
CHa [(n+1)T] = Ha[(n+1)T] Ha[nT]

(2)

Representative of conditions of air in variable

humidity ambient and this change used to maximum of
membership function, with appropriate rule in a table lookup scheme, shown table 3. For other step will be to adjust at
table 1. and adapt for schedule drying at figure 1

Table 3. Look-up table for membership Function for humidity of ambient

Ha
Hot
Rather-Hots
Swarm
CHa
Rather-Hot
-H
Hot
Hot
Rather-Hot
Swarm
-M
Hot
Rather-Hot
Swarm
-S
Hot
Rather-Hot
Swarm
Z
Hot
Rather-Hot
Swarm
+S
Hot
Rather-Hot
Swarm
+M
Hot
Rather-Hot
Swarm
Fresh
+H

Implementation Control System

Automatic control unit has built-in programs for
drying of all wood types regardless of a starting wood
moisture content. It controls drying process automatically,
so the operator presence is not needed during drying.
Larger energy consumption is only during first day during
heating stage when electrical heaters are turned on until

ISSN: 2085-6350

Fresh

Cold

Swarm
Fresh
Fresh
Fresh
Fresh
Fresh
Cold

Sejuk
Cold
Cold
Cold
Cold
Cold
Cold

working temperature is reached. Later, during drying stage

these heaters are rarely turn on.
If a boiler installation exists in the drying complex, it makes
sense to mount heat exchangers for hot water or steam
(depending on the boiler) in the kiln. In that case, energy
from the boiler would be used for heating of the drying
chamber. That way kiln becomes combined (automatically
uses boiler heat). This option is interesting only for a large

Conference on Information Technology and Electrical Engineering (CITEE)

21

Proceedings of CITEE, August 4, 2009

capacity kilns and in the cases where electrical power

supply is unstable (large voltage or current oscillation or
frequent power cut-downs during winter periods). Drying is
in that case performed in conventional way, heat pump is
turned-off, and electrical energy is used only for flow fans.
Combined kiln unites good characteristics of both
dehumidifying and conventional kiln. That provides the
greatest drying quality with the least energy consumption
for the shortest time.
Drying quality in dehumidifying kilns is
significantly better than in any conventional kiln, because
the drying is equable in whole timber load, so there is no

danger of developing degrade in the timber (cracks, checks,

warping...). drying is fast enough since the working
temperature can reach up 60 0C to. These kilns are
environment friendly because all wastes that are produced
during drying process are harmless for water and
environment. Especially good results are achieved with
drying of natural (untreated) wood (beech, ash,...).
Input variable in measurement by SHT11-sensor
and shown in figure 5. Schematic control system in figure
6 and flowchart fuzzy controlller illustrated at figure 7. In
table 4 describe input variable. and output variable used to
fuzzy rule suram shown at table 5

sprayer

Display 7 segmen
Damper in

fan

Motor Stipper
on/off

heater

Relay

Glass
DISPLAY

Td
Damper out

Microcontroller
AVR Atmega 128

RTC

Ta
Ha

Rd

Figure.5. Control system for prototyping wood drying kiln

Microcontroller AVR Atmega 128 have a capacity

a big amount 128k flash, 53 pin I/O, 6 channel PWM and 8
channel ADC 10-bit; then to used for application system
control complex. Microcontroller AVR Atmega 128 is a
microcontroller AVR Atmel 8 bit family, with specification
are :

128 Kb Flash PEROM

4Kb EEPROM, 4Kb SRAM

On Chip Analog Comparator

8 Channel 10 bit ADC

2 8 bit PWM
6 PWM with resolution programmable (2-16 bit)

Dual Programmable UART

SPI Interface
Programmable Watchdog with On Chip Oscillator
Adjustable VREF ADC
53 bit I/O
Power On Reset dan Programmable Brown out
detection
Internal Calibrated RC Oscillator
The pocket included the cabel ISP, CD and cabel
RS232

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

Proceedings of CITEE, August 4, 2009

22

Figure.6. Schematic fuzzy controller system

Tabel 4. Input variable
No.
Variable
Temperature Drying Td
1.
Temperature Ambient Ta
2.
Humidity Drying
Rd
3.
Humidity Ambient
Ha
4.

Range
0 150 0C
0 150 0C
0 100 %
0 100 %

Describe
Weather
Conditions of air

Tabel 5. Output variable and implemented the fuzzy rule suram

Actuator
No.
Rule
Conditions
Heater Damper Sprayer
To drop of Temperature Drying Td and to hoist
SUR-AM 1
off
off
S
of Humidity Drying Rd by very suddenly
1.
[Heating- Process]
To hoist of Temperature Drying
by very
SUR-AM 2
H
off
off
2.
suddenly and to drop of humidity drying Rd.
To hoist of Temperature Drying Td and to stay of
SUR-AM 3
H
D3
off
3.
Humidity Drying Rd.
To hoist of Temperature Drying and to drop of
SUR-AM 4
H
D2
S
4.
humidity drying Rd.
To drop of Temperature Drying Td and to stay of
SUR-AM 5
H1
D3
off
5.
Humidity Drying Rd
To drop of Temperature Drying Td and to drop
SUR-AM 6
H2
off
S
6.
of Humidity Drying Rd
To stay of Temperature Drying Td and to stay of
SUR-AM 7
off
off
off
7.
Humidity Drying Rd
To stay of Temperature Drying Td and to adjust
SUR-AM 8
H
D1
off
Humidity Drying Rd with Humidity ambient Ha
8.
[Equalizing-Process]
Note: D1
: Damper ON : 5 menit
H
: Heater ON : 15 menit
D2
: Damper ON : 2 x 1 menit
H1
: Heater ON : 10 menit
D3
: Damper ON : 1 x 1 menit
H2
: Heater ON : 5 menit
S
: Sprayer ON : 1 x 1 menit.

ISSN: 2085-6350

Conference on Information Technology and Electrical Engineering (CITEE)

23

Proceedings of CITEE, August 4, 2009

Figure.7. Flowcart fuzzy controller system

Result of Implementation the fuzzy rule suram for
measurement of temperature drying for schedule Albasia
Albizia wood drying, with specification : wood dimension :

4 x 6 cm, amount : 22 item. The figure 8 and figure 9.

shown , that measurement of temperature drying can be to
follow schedule drying.

Measuremet of Temperature Drying and MC-wood

7:26

4:02

0:38

21:14

17:50

14:26

11:02

7:38

4:14

0:50

21:26

18:02

14:38

11:14

7:50

4:26

1:02

21:38

18:14

14:50

11:26

14:47

11:23

M
CW
ood(%
)

Tem
peratureD
rying(oC
)

3-7 Juni 2008

70
60
50
40
30
20
10
0

time

Temp.DryingMCWoodSetpointTd
Figure 8. Measurement of temperature drying and MC-Wood at 3-7 Juni 2008

Measurement of Temperature Drying and MC-Wood

10-14 Juni 2008
60
50
40
30
20

5:28

2:46

0:04

21:22

18:40

15:58

13:16

10:34

7:52

5:10

2:28

23:46

21:04

18:22

15:40

12:58

10:16

7:34

4:52

2:10

23:28

20:46

18:04

15:22

12:40

9:58

7:16

4:34

1:52

23:10

20:28

17:46

15:04

9:40

10
12:22

M
CW
ood(%
)

Tem
peratureD
rying(oC
)

70

Time

Temp.DryingMCWoodSetpointTd
Figure 9. Measurement of temperature drying and MC-Wood at 10-14 Juni 2008

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

Proceedings of CITEE, August 4, 2009

24

Conclusion
Rule suram based of fuzzy logic with variables of
weather is temperature ambient and conditions of air is
humidity ambient, and it can implemented for wood drying
process. The membership function of variable of state
represented in error value and change error with typical of
triangle and
trapezium. Result from analysis and
evaluation, the 8 fuzzy rule to control for output system can
be constructed in a number of way of weather and
conditions of air. It used to minimum of the consumption of
electric energy by heater. The rule suram used to stability
and equilibrium of schedule of drying in chamber by control
of temperature and humidity. The result of implemented of
fuzzy rule suram with the modification of membership
function in range [0.5, 1] represented approximate to he
conditions riel are near enough and the actuator doing in the
conditions data riel
Reference
[1]. Nogueira, A., et all, Simulation and control strategies
for an energetically efficient wood drying process,
EFITA/WCCA Joint Congress on IT in Agriculture,
Vila Real Portugal, page 244 -251. 2005
[2]. Lim P.K.and Natalie R.S.,; Climate control of a storage
chamber using fuzzy logic; Proceedings of the 2nd
ANNES IEEE, 1995.
[3]. Virginia Tech, Virginia Cooperative Extension,
Virginia State University,, Design Of Solar-Heated
Dry Kiln, 2007
[4]. Kantay, R., Oner Unsal, Suleyman Korkut, 2002,
Drying problems of fast growing tree species :
Evaluation of Maritime pine (Pinus pinaster Alt) and
Eucalyptus (Eucalyptus camaldulensis Dehn) wood.,
Proceedings
International
IUFRO
Meeting
Management
of
fast
Growing
Plantations,
DIV.4.04.06,pp 208-212, 11th-13th September 2002,
Izmit-Turkey
[5]. Wang L.X, A course in Fuzzy Systems and Control,
Prentice-Hall International, Inc, 1997.
[6]. Joseph C.G and Gary D.R, Expert system Principle and
Programming, Thomson Course Technoloy, Fourth
Edition, Canada, 2005.
[7]. Klir, G.J, and Bo Yuan, Fuzzy sets and Fuzzy Logic
Theory and Applications, , Prentice-Hall International,
Inc, New York, 1995.

[8]. Duenas, A., and Dobrila P, A new Approach to Multi

Objective Single-Machine Scheduling Problem under
Fuzzines, Decision Support in an uncertain and
complex word: the IPIP TC8/WG8.3. International
Conference, 2004
[9]. Haque.M.N., Modelling of Solar Kilns and The
Development of An Optimised Schedule for Drying
Hardwood Timber, Thesis Ph.D., Department of
Chemical Engineering, University of Sydney., 354p.
2002
[10]. Dion,J.M., L.Dugard, A.Pranco, N.M. Tri,
J.W.Horwood., MIMO Adaftive Constrained
Predictive Control Case Study : An Environmental
Test Chamber, Automatica, 27, 1991, 611- 626,
Pergamon-Press
[11] Garrahan.P., and Jeff Kidlark, 2001, Drying
Hardwoods - Northern Hardwood Initiative : Industry
Resource Manual, Forintek Canada Corp, 2665 East
Mall, Vancauver, B.C. V6I 1W5.
[12] Sun, Z.F., C.G.Carrington., C.Davis, Q.Sun, and
S.Pang, Drying Radiata Pine Timber under
Dehumidifier Conditions : Comparison Of Modelled
Results with Experimental Results, 8th nternational
IUFRO Wood Drying Conference-2003 page 39 44
[13]. Laurenzi. W., V.Popa, Gh.Comsa, Computer Aided
Design Of Drying Schedules In Order To Control
The Drying Process, 8th International IUFRO Wood
Drying Conference, 2003, 302 307.
[14] Nogueira, A., et all, 2005, Simulation and control
strategies for an energetically efficient wood drying
process, EFITA/WCCA Joint Congress on IT in
Agriculture, Vila Real Portugal, 2005, 244 -251
[15] Skuratov. N.V., Computer Simulation and Dry Kiln
Control, 8th International IUFRO Wood Drying
Conference, 2003, 406 412
[16] Wang, X.G. Liu, W. Gu, L. Sun, C.J. Gu, C.E.
de Silva, C.W., Development of An Intelligent
Control System for wood drying proceeses ,
Advanced Intelligent Mechatronics Proceedings.
2001 IEEE/ASME International Conference. Vol.I,
2001, 371 376.
[17] Situmorang, Z., et all, The Schedule Of Optimal
Fuzzy Controller Gain With Multi Model Concept
for a Solar Energy wood Drying Process Kiln,
International Journal Of Operations and Quantitative
Management, 2009..

ISSN: 2085-6350

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

25

OPTIMAL CAPACITOR BANK LOCATION IN THE PRIMARY FEEDER

WITH TYPICAL FLAT LOAD
Hermagasantos Zein
Department of Energy Conversion Technology, Bandung State Polytechnic
Mail Box: Pos 1234, Indonesia
Email: hermaga_s @ Yahoo.co.id
Mobile phone: 081321108991

Abstract- A capacitor bank is not only used for

improving power factor, but can also be used to minimize
losses in electric system networks. This paper is aimed to
determine an optimal location of capacitor bank in the
typical flat load of a primary feeder in order to get maximal
losses-saving. The capacitor bank position will affect losssaving depending on the number of nodes and the
magnitude of compensation current generated by the
capacitor bank. By simulation, the optimal position of
capacitor bank from the substation is further than the center
of feeder. For 100% compensation and unlimited number of
nodes, the optimal location is 50% of feeder length. Optimal
condition for 15 nodes, for 100% compensation is 0.53 pu
and 30% compensation is 0.92 pu from the substation.
Generally, the maximal saving occurs between 50% and
70% compensation, and the location of the capacitor bank
unchanged for the number of nodes greater than 15. Then,
the optimal locations for multi capacitor banks also depends
on both the number of nodes and the magnitude of
compensation current generated by every capacitor bank.
For five capacitor banks with the number of nodes greater
than 10 nodes indicate that the optimal locations of them are
not changed for 15% compensation, i.e. 0.3 pu, 0.5 pu, 0.7
pu, 0.8 pu and 1.0 pu from the substation.
Keywords- bank capacitor, losses saving, primary feeder,
optimal location
I. INTRODUCTION
An inductive load in the primary feeder will decrease
the power factor of system because it has to supply reactive
power to the load. For feeders with low power factors (less
than 89%) have to be corrected until the reactive supply
become smaller. If the compensation is aimed only for
correcting of the power factor of system, the capacitor bank
location will be independently (no problem). Another
problem appears while correcting the power factor, how
much losses in the feeder can be reduced by installing of
capacitor bank. It is possible because of the capacitor bank
current will reduce reactive current produced by the load (as
dominant contribution), so, the current flowing and losses in
the feeder can be reduced. Significance of reducing in losses
in the feeder is influenced by the location where the
capacitor bank is installed. If losses reducing as consider in
installation of capacitor bank in the feeder, the capacitor
bank location have to be looked for in order to reduce losses

to be maximum. Where the losses in the distribution system

(primary feeder is dominant) is greater than 60% from total
losses system.
Figure-1 is a primary feeder with lumped load and
current (I) flowing in the feeder. Supposed that resistance of
the feeder is R, power factor ( Fb = Cos ) lagging and
reactance of the feeder is ignored, then, the current flowing
in the feeder is also lagging with respect to the voltage by
phase angle . The current can be formulated as

I = ICos jISin
= I a jI r
where: I a = ICos is an active current
I r = ISin is a reactive current

(1)

Figure-1: The feeder with lumped load

From figure-1, losses can be derived from the equation-1
and yields:

Pr = ( I a2 + I r2 ) R

(2)

If at the load is installed the capacitor bank that yield

current, Ic, losses in the feeder can be determined from the
equation-2 and yield

Prc = [ I a2 + ( I r I c ) 2 ]R
= [ I a2 + I r2 2 I r I c + I c2 ]R

(3)

From the last two equations, the saving power by installing

capacitor bank is

Pr = Pr Prc
= 2 I r I c I c2

(4)

From equation-4, the losses which can be saved and depend

only on the reactive current. The active current is not shared
in this case.

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

26

Proceedings of CITEE, August 4, 2009

II. LOSSES REDUCTION BY THE CAPACITOR BANK

f ( xm ) =

Figure-2 is the reactive current curve in the feeder.

This curve is expressed as a function of distance (x) as:
I x= f (x)
(5)
Then, the losses can be counted in every segment with
distance dx from the feeder in the figure-2 as:

Prr = 3[ f ( x)]2

R
x'
L

(6)

Figure-2: Reactive current versus distance in

the feeder

R
x'
L

(7)

= 3RI c [2(a + bx) I c ]x

0

= 3RI c [(2a I c ) x + bx 2 ]

(8)

2a I c
b
x + x2

IT
IT
2
And let the losses of one unit is 3RI T and k = I c / I T ,
= 3RI T2

Ic
IT

whereas I T = nI , a = I T and b = ( I t I ) / L , the power

can be saved (losses saving) in the unit is

Prr ( pu ) = k[(2 k ) x

The last two equations is able to reduce losses in unit

distance as:
(9)

where: dx=dx/L
From equation-9 can be determined the optimal location of
the capacitor bank through derivative of the equition-9 to
distance and the result is an optimal location (xm) as:

Prr
= 3RI c [2 f ( xm ) I c ] = 0
x

From equation-8 and equation-9, losses which can be saved

depend on the location of a capacitor bank along the feeder.
Let the location of a capacitor bank is x-distance from the
substation, the losses can be determined as:
x

Prrc =

Figure-3 is a feeder with the typical flat loads (10

nodes). The mathematical model of current function along
the feeder is linear.
f ( x) = a + bx
(12)

From figure-2, let x is the location of a unit capacitor bank

from the substation and Ic is phase current of the bank, now
the total losses is:

Prr = P rr Prrc = 3 RI c [ 2 f ( x ) I c ]x

III. FEEDER WITH TYPICAL FLAT LOADS

Prr = P rr Prrc = 3RI c [2 f ( x) I c ]x

where: Prr is total losses affected by reactive current.

f(x) is reactive current function along feeder
R is total resistance of feeder.
L is total distance of feeder.
dx is the distance element.

3R x
3R L
[ f ( x) Ic ]2 x' + f 2 ( x)x'
L 0
L x
3R x 2
3R L
=
[ f ( x) 2Ic f ( x) + Ic2 ]x'+ f 2 ( x)x'
L 0
L x
3R L 2
3R x
=
f ( x) + I c [ Ic 2 f ( x)]x'
L 0
L 0

(11)

Figure-3: A Feeder with typical flat loads

Therefore, total losses along feeder can be counted by

integrating the equation-6 and resulting:

Prr = 3 [ f ( x)]2

Ic
2

(10)

n 1 2
x ]
n

(13)

For n=10, plots of equation-13 to k variation are

presented in the figure-4. The figure points out that
maximum saving of 0.33 pu accurs at k=0.7 (or capacitor
bank compensation 70%) by distance x=0.7 pu. Whereas
compensation 100%, maximum saving is 0.28 pu and the
location between x=0.5 pu and x=0.6 pu. The optimal point
goes down by increasing of of distance from x=0.5 pu to
x=1.0 pu. Whereas the saved maximum power can be
increased by reducing the compensation downs to 60% and
it goes down again if the compensation ins increased. From
the curve in the figure-4, the best compensation are at
k=0.5, k=0.6 and k=0.7.

This equation can be simplified further as:

ISSN: 2085-6350

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

27

Losses saving (N=10)

0.35

k=0.5k=0.4

k=0.6
k=0.7
k=0.8

k=0.3

k=0.9

0.30

0.25

0.20

n
k n 1 n
k
(1 )
(1 )
Prrmak ( pu) = k (2 k )

n
1
2
n
n
1
2

= k[(2 k )

k=0.2

n
n
(2 k ) 2 0.25
(2 k ) 2 ]
n 1
n 1
n
= 0.25k
(2 k ) 2
n 1

k=1,0

k=0.1
0.05

n 1
]
n

= k[0.5

0.15

0.10

(15)

Losses saving
0.6

0.00
0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0.5

Jarak dari GI (x)

0.4

Figure-4: Losses saving curve versus

variation of capacitor bank
location for n=10

k=1.0

0.3

k=0.3
0.2

k=0.1

0.1

0.0

IV. OPTIMAL LOCATION

From equation-11, the optimal location of the
capacitor bank from the substation, xm, for flat load (figure3) can be determined as:
2a I c
n
I
n
k
(14)
1 c =
xm =
=
(1 )
2b
n 1 2IT n 1
2
where: k = I c / I T is capacitor current and reactive current
ratio at supply.
n =2,3 is number of loads in the feeder.
The equation-14 points out that optimal location of a
capacitor bank depends on the number of loads or nodes, n,
and compensation capacity, k. Figure-5 is curves of optimal
location for difference n and k. For the capacity of k,
optimal distance becomes smaller when n is increasing at
some extend then it becomes constant for further increasing
of n. The optimal location x declines sharply for n less than
10 and for n greater than 10 nodes the decline is not
significant (close to constant). Example for k=1 pu, x=0.556
pu and it will lead to x=0.5 pu on n = and k=0.8 can be
result x= 0.667 and it will lead to x=0.6 at n = .
x
1.0

k=0,1
k=0,2

0.9

k=0,3
k=0,4

0.8

k=0,5
k=0,6

0.7

k=0,7
k=0,8

0.6
k=0,9
k=1.0

0.5
2

10

11

12

13

14

15

16

17

18

19

Figure- 5: Optimal location of a capacitor bank

in the flat load feeder
Then amount of saving that is gotten for optimal
location versus the number of nodes variation can be
derived through equation-13 and equation-14, and it yields
equation-15. It pointed that maximal saving only depend on
n and k. Plot of the equation-15 is presented in the figure-6.
It states that the influence of optimal location versus
number of nodes for variation of compensation. It declines
sharply from n=2 to n=10, and whereas for n>10 there is
nearly no change in loss saving. Like that has been
explained before, the good savings are for k=0.5 to k=0.7.

9 10 11 12 13 14 15 16 17 18 19 20

Figure-6: Amount of power can be saved for

a capacitor bank location versus n

V. FOR M-CAPACITOR BANK

Capacitor bank can be put in the feeder more than
one unit. By assumption that m-capacitor banks are
identically, the losses as consequence of the m-capacitor
bank can be written as:
x2
3R x1
[ f ( x) mIc ]2 x' + [ f ( x) (m 1) Ic ]2 x' + ...
x1
L 0
3R L
+ f 2 ( x)x'
L xm
3R x1 2
=
[ f ( x) 2mIc f ( x) + m2 Ic2 ]x'+
L 0
3R x 2 2
3R L
[ f ( x) 2(m 1) I c f ( x) + (m 1)2 I c2 ]x'+... + f 2 ( x)x'
L x1
L x
3R L 2
3R x1 2
=
f ( x) +
I c [m I c 2mf ( x)]x' +
L 0
L 0
3R x 2
3R x
I c [(m 1)2 Ic 2(m 1) f ( x)]x' + + I c [ Ic 2 f ( x)]x'
L x1
L 0

Prrc =

The saving can be counted such as a single unit that was

explained before and it results:
x1

Prr = 3RI c [2mf ( x) m 2 I c ]x +

0

x2

3RI c [2(m 1) f ( x) (m 1) 2 I c ]x + ...

x1

+ 3RI c [2 f ( x) I c ]x
m

x1

= 3RIc { [2mf ( x) m 2 I c ]x
0

x1

x2

x2

[2(m 1) f ( x) (m 1) 2 I c ]x +

[2(m 1) f ( x) (m 1) 2 I c ]x
[2(m 2) f ( x) (m 2) 2 I c ]x + ...}

From the last equation , it can be simplified in term of unit

as equation-16. This equation points out that the saving is
also influenced by location of the capacitor bank-i.

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

28

Proceedings of CITEE, August 4, 2009

Prr =

Ic
IT

m xm+1i

i=1

2 f ( xm+1i ) (2i 1)IC

xm+1i
IT

(16)

VI. OPTIMAL LOCATION OF M-CAPACITOR BANK

Optimal location each capacitor bank can be found
by of the equation-16 resulting:

2 f ( xm+1i ) (2i 1) I C = 0

(17)

Substitution the current function, f(x), then the optimal

location each capacitor bank can be determined as loads in
the equation-18.

xm+1i =

n
2i 1
k)
(1
n 1
2

(18)

Equation-18 explains that the optimal location of

each capacitor bank depends only on the number of nodes
and capacity of compensation applied. Table-1 gives results
for 5 capacitor banks with compensation 15% each others
from reactive current supply. For n>15 optimal location of
each capacitor bank is constant. Significant changes occur
between n=2 to n=6, for n=2: 4 capacitor bank at x=1.0 pu
and one capacitor bank at x=0.7 pu. When it is compared to
n=6: only one capacitor bank at x=1.0 pu and the others
distributed in the feeder.
Table-1: Optimal location of 5 capacitor bank in the
feeder for k=0.15 with variation of number of
nodes
Number
of nodes

Load
current

Location of five capacitor

banks

I(pu)

x1

x2

x3

0.50

0.7

1.0

0.33

0.5

0.7

0.25

0.4

0.20

1.0

x4
1.0

x5
1.0

0.9

1.0

1.0

0.6

0.8

1.0

1.0

0.4

0.6

0.8

1.0

1.0

0.17

0.4

0.6

0.8

0.9

1.0

0.14

0.4

0.6

0.7

0.9

1.0

0.13

0.4

0.5

0.7

0.9

1.0

0.11

0.4

0.5

0.7

0.9

1.0

10

0.10

0.4

0.5

0.7

0.9

1.0

11

0.09

0.4

0.5

0.7

0.9

1.0

12

0.08

0.4

0.5

0.7

0.8

1.0

13

0.08

0.4

0.5

0.7

0.8

1.0

14

0.07

0.4

0.5

0.7

0.8

1.0

15

0.07

0.3

0.5

0.7

0.8

1.0

16

0.06

0.3

0.5

0.7

0.8

1.0

17

0.06

0.3

0.5

0.7

0.8

1.0

18

0.06

0.3

0.5

0.7

0.8

1.0

19

0.05

0.3

0.5

0.7

0.8

1.0

20

0.05

0.3

0.5

0.7

0.8

1.0

ISSN: 2085-6350

VII. ANALYSIS
Current in the flat load feeder (figure-3) declines
linearly by the increase of distance from substation to the
end of feeder. The capacitor bank put in the flat load feeder
is not only improving the power factor but also saving
energy. The saving depends on location as consequence of
the current change. The results of installing a capacitor bank
states that maximal saving occur between compensation
50% to 70%, as presented in figure-4. Whereas optimal
location varies with number of nodes and the compensation
capacity applied. For indefinite number of nodes with
compensation 100%, optimal location is at the center of
feeder, it is optimal location with the shortest distance from
substation. If n is reduced, the optimal location will become
longer from half distance of feeder. Figure-5 points out that
for n>15, the optimal location is nearly constant and it is
found the same condition as is pointed by figure-6. For mcapacitor bank, optimal location of each capacitor banks
depends on its compensation capacity. Table-1 points out
that for n>10 nodes, optimal location each capacitor banks
is close to constant. Whereas for n<5, lumped all of
capacitor bank is occurred at the end of feeder (for n=5 is
two capacitor banks and n=2 is four capacitor banks)

VIII. CONCLUSION
Capacitor bank position in feeder is truly determining
saving which depends on the number of nodes or load
location and capacity of current compensation produced the
capacitor bank. The result in figure-4 shows than in flat load
of feeder, the optimal location of a capacitor bank is in the
center by unlimited number of nodes and 100%
compensation. Whereas small compensation or equal
reactive load current, the optimal location is at the end of
feeder. So location of single capacitor bank will be between
half distances to the end of feeder. Maximal saving of single
capacitor bank with flat load is occurred on between
compensation 50% and 70%.
Optimal location of single capacitor bank in the flat
load feeder is not changed for number of nodes greater than
14. Thus, the saving is truly influenced by compensation
capacity applied (figure-5), for full compensation (100%)
optimal location is 0.53 pu from substation, and with
compensation 30%, the location is at 0.92 pu from
substation.
The optimal location for m-capacitor bank depends
also on the number of nodes and compensation capacity
applied. By simulation, for 5 capacitor bank with
compensation 15% shows the optimal location of the fifth
capacitor bank does not change for number of nodes greater
than 10 nodes, that is at 0.3 pu, 0.5 pu, 0.7 pu and 1.0 pu,
repeatedly from substation.

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

IX. SUGGESTION
In the improving power factor it will be very
valuable to consider saving capacity. Before improving
the power factor of feeder it needs to evaluate how much
benefit that can be earned from both saving by limiting
reactive cost and reducing of power supply, where
equipment capacity become loose.
REFERENCES
[1] Gonen, T., Electrical Power Distribution System
Engineering, McGrow-Hill,Inc., Copyright, 1986.

29

[2] Puloka T., Simulation of Resonance in a Small-Scale

Network Laboratory, Proceeding Power on,
International Conference on Vol.3., 2000.
[3] Zein, H., Transmission and Distributions, Lecture
Dictate, Department of Energy Conversion
Technology, Bandung State Polytechnic, 2008.
[4] Zein, H. Simulasi kinerja jaringan transmisi dalam
kemampuannya menyalurkan daya dari suatu
pembangkit, Prosiding ISSN:1979-911X, IST
AKPRIND, Yogyakarta, 2008.
[5] Zein, H., (Perkiraan Pasokan Daya Sistem JawaMadura_Bali sampai Tahun 2016 Berdasarkan Index
LOLP Satu Hari Pertahun, Majalah Ilmiah
TRANSMISI, Teknik Elektro Undip, Semarang,
2009.

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

30

Proceedings of CITEE, August 4, 2009

DUAL PARALLEL POWER CONVERSION CONVERTER SUPPLIED BY PHOTOVOLTAIC FOR BASE

TRANSCEIVER STATION (BTS) POWER SUPPLY
1. Kartono Wijayanto
Politeknik Negeri Bandung
E-mail address: [email protected]
2. Yanuarsyah Haroen
School of Electrical Engineering and Informatics ITB.
E-mail address: [email protected]

ABSTRACT
For the reliability reason Base Transceiver Stations (BTS)
are located in the rural area or as backbone station need the
second or the third power supply beside PLN and the diesel
generator. Photovoltaic system used to supply the BTS that
operates at the voltage of 48 volt, 2000 watt, normally
equipped by backup battery. By such voltage and power,
needs 24 photovoltaic modules composed in two arrays
comprise of 2 x 6 photovoltaic for each. Each photovoltaic
has capacity Pmax=150 watt and Vmax=34.5 volt. In this
paper, the two photovoltaic (in series connection) are using
as the modeling of the photovoltaic system. When some
parts of the photovoltaic surfaces are shadowed by cloud, it
will cause the I-V characteristics change and have two
maximum power areas. In this case, it will cause MPPT
with single converter will not be able to extract the
maximum power. To overcome that case, a new topology,
Dual Parallel Power Conversion Converter modeled for the
power supply of BTS. The converter consists of series
connected of dc-dc positive and negative polarity converter
where each converter controlled independently for MPPT.
As a controller, microcontroller produces two PWM (Pulse
Width Modulation) signals. The value of the duty cycle D
determined by MPPT algorithm using standard modified
method. The experiment has done using two photovoltaic
simulators and a microcontroller ATMEGA 16. The
performances of dual converter parallel power conversion
have good results and it is capable to maintain the operation
of photovoltaic at its maximum area and by modified
standard method, the time response is faster than the
standard method.

exhausting of fossil where the reserves are getting more

diminish. Indonesia situated on the equator, has abundant
potency of sun energy with the average intensity of 4.8
Kwh/m2/day or equal to 400 W/m2. The Base Transceiver
Station as a backbone station or others functioned as up and
down link communication station gets limited covered area.
In the rural areas, the supply of electricity from PLN often
gets disruption or black out for such a long time. This
condition will increase the consumption of fuel and the
operational cost of BTS becomes more expensive.

II. PHOTOVOLTAIC
Photovoltaic is a cell consists of p-n junction. When its
surface exposed to the sun light, it produces a direct current
(DC) voltage about 0.5 Volt. In order to get a bigger voltage
or current capacity, numbers of PV are connection in series
or parallel. The simplest equivalent circuit of a PV cell is a
current source in parallel by a diode, as shown in Fig. 1.
The properties of the PV current explained as follow:

q
I = I ph I OS exp
(V + IRs ) 1 V + IRs
nkT
RSH

(1)

Keywords : Photovoltaic, MPPT, dual converter paralel

power conversion, microcontroller
Fig. 1: Equivalent circuit diagram of PV

I. INTRODUCTION

The power supply which uses photovoltaic is really

appropriate to be used particularly in the areas of not getting
energy supply from the State Electricity Enterprise (like
PLN) or even in the isolated areas as in the mountain range
areas, isolated islands and other areas where other sources
of energy available become not economic to develop. The
usage of photovoltaic can decrease the consumption of fuel
that is getting more expensive. It will also decrease the
ISSN: 2085-6350

I ph = I SCR + K I (T 25)
100

(2)

Where:
I = PV terminal current
V = PV terminal voltage
Iph = Current generated
q = Electron charge
k = Boltzman constant

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

T
Rs
Rsh
Iscr
Ios

31

= Temperature of PV
= Series resistance
= Parallel resistance
= Short circuit current at 250C, 1000 w/m2.
= Saturation reverse current of PV

The output of the current source is proportional to the light

falling on the cell. When the cell is short circuit, negligible
current flows in the diode and the current flows in the
external circuit. In open circuit, this current
shunted
internally with the intrinsic p-n junction diode. The
characteristic of I-V which has the variety of intensity
depicted in Fig 2a. In case of series connection photovoltaic
which the surface are shadowed partly by the cloud depicted
in Fig. 2b.

Fig 3: Topology of DPPCC

Vp1

P
1

Fig 2a

Vp2

Fig 2b

Fig 2a: I-V characteristic with variety of intensity

Fig 2b: Two series PVs partly shadow by cloud

Fig 4: Equivalent circuit of DPPCC at S1&S2 on

III. DUAL PARALLEL POWER CONVERSION

CONVERTER
The problem of this research is according to two
maximum regions as shown in Fig. 2b that solved by using
a new topology of converter, dual parallel power conversion
converter (DPPCC). The diagram of DPPCC depicted in Fig
3. At the steady state condition, the capacitor CL1 and CL2
are charging for a moment and as well the batteries Vb1 and
Vb2. When the switch S1 is on, the current, IC flows trough
the Inductor L and charges the battery Vb1. When the S1 is
in off condition, the current is flowing from the inductor L
to charge the capacitor CP1 and the battery Vb1. The
converter in the lower position is working under the same
principles with opposite current flow.

The inductor current is linearly increasing relative to the

time and achieved the maximum value IL. If tton=DT,
then:
(5)
Vc
I L1 = L1 DT + I L,0
L1
The peak ripple current is:

I L1 = I L1 - I L,o =

VcL1
DT
L1

(6)

When S1 and S2 in off condition, the equivalent circuit is

depicted in Fig 5, no energy flows from the photovoltaic.

Analysis of DPPCC
The equivalent circuit with S1 and S2 is in on condition, is
depicted in Fig. 4. With assuming the VcL1 is constant, the
equation is:

L1

diL1 (t )
+ RL1iL1 (t ) = VcL1
dt

(3)
Fig 5: Equivalent circuit of DPPCC at S1&S2 off

By neglected the value of RL, the inductor current is:

iL1 (t ) =

VcL1
t + I L,0
L1

(4)

The amount of energy of inductor L1 discharged and the

current ID3 is flowing to charge the capacitor CP1. As well as
the current ID4 flowing to charge the capacitor CP2.
The batteries Vb1 and Vb2 charged by the capacitor CP1 and
CP2 and the equations are:

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

32

Proceedings of CITEE, August 4, 2009

diL1 (t )
di (t )
= VcP1
dan L2 L 2 = VcP 2
dt
dt
(7)
VcP1
VcP 2
iL1 (t ) =
t + I L1,maks dan iL2 (t ) =
t + I L2 ,maks
L1
L2
L1

The inductor current is decreasing relatively to the time and

achieved the minimum value of IL. If ttoff = (1-D)T, then :
IL =

(VCP )
(1 D )T + I L ,max , where assumed Vc1 = Vc2
L
and L1 =L 2

(8)

The peak current ripple is:

I L = I L ,maks I L ,o

(V )
= CP (1 D)T
L

(9)

Fig 7: Capacitor, CP current

Fig 8 shows the capacitor current ICL that consist of current

at switch S in on condition and switch S in off condition and
the equations are:

VP VO
DT ,
where : = RC

V
= I L = O (1 D )T
2L

I CL ,maks =

(16)

I CL ,min

(17)

From equations above, we get:

D
D
VL atau VO =
Vi ,
1 D
1 D
where : VL = Vi

VCP = VB =

(10)

Vi (VO / 2)

I L

= VP - VB
VO = VB

Fig 8: Capacitor current waveform

The inductor current flowing depicted in Fig 6.

The inductor current at the boundaries shown in Fig 9

I L

Fig 6: Inductor current waveform

By using ideal components then we have:

VL I L = VB I B
D
VL )
= IB (
1 D
1 D
1 D
IB =
I L or IO =
Ii
D
D

Fig 9: Inductor current at boundaries

(11)

(12)

The maximum and minimum currents of capacitor CP are:

(V )
I L
I C p , m aks =
= O (1 D )T
(13)
2
2L
(V )
I L
(14)
= O (1 D )T
I C p ,m in =
2
2L
The capacitor current ICP waveform depicted in Fig 7 and
the voltage equation is:

VC =

(1 D)
Q (VO )
=
(1 D)T 2 =
(VO )
C 8LC
8LCf 2

ISSN: 2085-6350

(15)

VT
1
I LB = I L, p = B S D(1 D)
2
2L

(18)

With VB kept constant, the maximum of ILB has achieved

the value of D = 0.5 and the average of inductor current is:

I LB =

TS VO
TV
(1 D) dan I OB = S O (1 D) 2
2L
2L

(19)

IV OPERATION OF THE SYSTEM

The system consist of the photovoltaic, the DPPCC, and the
batteries as a back up or storage device that its supplied by
the DPPCC or its supplied by the other power supply like
Diesel Generator. Block diagram of the system depicted in
Fig 10. Table 1 shows the operation of the system.

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

33

Fig 10: Block diagram system

With application of the relays/contactors sw1, sw2, and

sw3, the usage of the sun energy would be maximal through
the operation of the relays according to table 1.

Fig 11: P&O modified Algorithm

Table 1: Logical operation of the system

1
2

SW
1
Off
Off

SW
2
Off
Off

SW
3
Off
On

Off

On

On

On

Off

On

On

On

On

No

Condition/Status
Battery low level
Battery Discharging
At night or low intensity
battery charging or
floating
At day w/ normal
intensity battery
charging or floating
At day w/ low intensity
battery charging or
floating

From the table 1, Maximum Power Point Tracker (MPPT)

applied to the DPPCC adjusts the duty cycle D
automatically, so the operation point of the photovoltaic is
in the maximum power region and the maximum energy
extracted. The algorithm of MPPT in this research is using
the modified standard P&O. The MPPT tracks the I-V
characteristic of photovoltaic, the current and voltage
sensed and the calculation of the power. The power Pnow
compared with the power past Ppast. If the power now
greater than power past and if the curve still at the linear
part, the duty cycle D increase significantly and if the power
now less than the power past then the duty cycle D
decreased by a little. The MPPT modified algorithm shown
in Fig 11. The MPPT algorithm is part of the operation of
the system. When the sun intensity low or the battery is full
or floating charged, the MPPT automatically generates the
duty cycle that matches to the load, but it is not the
maximum duty cycle. The algorithm of the system depicted
in Fig 12. The operation of the switch is match with the
procedure of the battery charging (bulk, absorption and
floating charging).

Fig 12: Algorithm of the system

V. EXPERIMENT & RESULT

An experiment was done to validate the modeling of
the system, The experiment is consist in two parts, part one
is the DPPCC converter as depicted in Fig 13 and part two
is the control of the system by using an ATMEGA 16
microcontroller as depicted in Fig 14.

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

34

Proceedings of CITEE, August 4, 2009

around the maximum value that shows the work of the

tracking algorithm.

Fig 13: Diagram of the system

In the experiment, two photovoltaic simulators are using,

with difference characteristic of sun intensity to test the
work of MPPT with new topology DPPCC. The sensing of
voltage is used voltage divider and used inverting amplifier
for negative voltage. The current sensing is used a 0.1 ohm
resistor inserted to the circuit and an inverting amplifier is
used to amplify the value to match to the microcontroller
input.

Fig 16: Photovoltaic current at the sun intensity change

Fig. 17 shows the performance of MPPT. The tracking

algorithm causes the controller tracks the I-V characteristic
(dots in line) and oscillates around the maximum power
point and the dots intercept the line, shows the I-V response
upon change of the sun.

Fig 16: Performance of the MPPT

Fig 14: Control system diagram

VI. CONCLUSION

The measurement result of DPPCC according to the

frequency of duty cycle D 4.5 KHz shown in Fig 15.

1.

2.

1>

1>

1) Ch 2:

1) Ch 2:

10 Volt 50 us

20 Volt 50 us

3.

1>

1>
1

4.
1) Ch 2:

10 Volt 50 us

1) Ch 2:

10 Volt 50 us

Fig 15: Current waveform of:

- Battery (left up) and capacitor (right up),
- Inductor (left down), photovoltaic (right down).

Fig 16 shows the photovoltaic current at sampling two

seconds. It shows, at the beginning of start, the duty cycle D
causes the PV current increases rapidly. At the maximum
current, the duty cycle D causes the PV current oscillates at

ISSN: 2085-6350

The Batteries supplied by the DPPCC charged

independently by a photovoltaic module as the
switches are on. They charged simultaneously
when the switches are off.
Each converter running with individual MPPT, that
the energy extracted from the PV independent each
other so the problem of two maximum regions to
be overcome.
The modified standard algorithm was work better
than the standard (speed and stability improved)
The application of the control method could also
be used to the other converter with bigger capacity.

REFERENCE :
[1]

DB Snyman & JHR Enslin.(1989), Combined Low

Cost, High Efficient, Peak Power Tracker And
Regulator For PV Applications CH2721-9/89/0000
067.

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

[2]

D.P.Hohm, M.E. Ropp.(2000), Comparative Study

of Maximum Power Point Tracking Algorithms Using
an Experimental, Progammable, MaxmumPower
Point Test Bed, IEEE 0-7803-5772-8/00.

[3]

Hirotaka Koizumi, Tamaki Mizuno, Takashi Kaito,

Yukihisa Noda, Norio Goshima, Manabu Kawasaki,
Ken Nagasaka, and Kosuke Kurokawa.(2006),
A Novel Microcontroller for Grid-connected,
Photovoltaic Systems, IEEE Transactions on
Industrial Electronics, Vol. 53, NO. 6, Desember
2006.

[4]

[5]

[6]

K. Hirachi, T. Mii, T. Nakashiba, KGD Laknath, M.

Nakaoka.(1996),
Utility-Interactive
MultiFunctional
for
Solar
Photovoltaic
Power
Conditioner with Energy Storage Batteries, IEEE 07803-2775-6/96.
Mohan, N., et al.(1995), Power Electronics
Converter, Applications, and Design, 2nd Edition.
John Wiley&Sons Inc. Canada.
Peter Kremer, Dipl Ing. (2001), Photovoltaic
Hybrid System Enhance Realibility of Power
Supply, 17th European photovoltaic energy
conference & exibition, munich.

35

[7]

Tao Yep Kim, Ho-Gyun Ahn, Seun-Kyu Park, YounKyu Lee.(2001), A Novel Maximum Power Point
Tracking Control For Photovoltaic Power System
Under Rapidly Changing Solar Radiation, IEEE 07803-2/01.

[8]

T L Pryor and N A Wilmot.(2004), The Effect of

PV Array Size and Battery Size on the Economics of
PV/Diesel/Battery Hybrid RAPS Systems , Murdoch
University Energy Research Institute.

[9]

Yanuarsyah Haroen, PROF. DR. IR. (2005), Diktat

Kuliah Elektronika Daya Lanjut, LPKEE ITB.

[10]

Yanuarsyah Haroen, DR, IR. (1983), Power

Conditioning of An Autonomous Microprocessor
Controlled System : Photovoltaic Generator
Maximum Power Point Tracking (MPPT) , INP
Toulouse.

[11]

Yukinori Kuwano, DR. (2008), The Future of Photo

Voltaic Power Generation, Photovoltaic Power
Generation Technology Reasearch Assosiation.

[12]

http://www.geocities.com/markal_bppt/publish/pltkcl
/plrahard.pdf, Analisis Potensi Pembangkit Listrik
Tenaga Surya Di Indonesia

[13]

http://www.iea-pvps.org, Trends in Photovoltaic

Application in Report country

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

36

Proceedings of CITEE, August 4, 2009

Integrated Buck Converter and Wind Turbine

Control System Medium Scale (100 W)
For Optimization Wind Power and Electricity
Power
Ali Musyafa, Soedibjo
Jurusan Teknik Elektro - Fakultas Teknologi Industri
Institut Teknologi Sepuluh Nopember ,Kampus ITS, Keputih Sukolilo, Surabaya 60111 Indonesia
Telp.: +62-31-5947188; + 62-31-5967288; fax +62-31-5923626
E-mail addresses: [email protected] ; [email protected]
I Made Yulistiya Negara , Imam Robandi
Jurusan Teknik Elektro - Fakultas Teknologi Industri
Institut Teknologi Sepuluh Nopember ,Kampus ITS, Keputih Sukolilo, Surabaya 60111 Indonesia
Telp.: +62-31-5947188; + 62-31-5967288; fax +62-31-5923626
E-mail addresses : [email protected] ; [email protected]

Abstract Discussion of the control system and power

electronics in wind turbine generation generally reviewed
separately, of so comprehensive and involves the study of
wind power. Electric power generators that are raised and
the output power or load to grid need to be more extensively
examined. Issues related to research performed to be wind
speed, wind speed modeling which is entries on wind turbine,
made further modeling and control design of a wind turbine
to get the optimal wind power, how to set up the play pitch of
the blade on a wind turbine. Wind power in the next feeder
induction generator as device the power of wind power into
electricity. Wind generated power value varies between (3040) DC voltage, while the output is the expected value (12-14)
DC voltage so that the necessary buck converter to
transformation ensure the output of electricity which is
determined. The modeling each block of system components
(wind turbine, generator, sensor, actuator and converter).
The designed the control system PID conventional and Fuzzy
logic control further applied to the system. By doing
simulations on system that has been integrated into the
system performance can be known; like rise time, settling
time, Overshoot, peak time and error steady state. From the
simulation shown that the integration of PID control and
converter produces a good response. But if compare with the
fuzzy logic control, Fuzzy logic control is a system that more
readily bring electricity better if composed with the PID
conventional control system.

Keywords - Wind turbine, Converter, PID conventional,

Fuzzy logic control, System performance.

I.

INTRODUCTION

Wind energy has increased interest in electrical power

applications, it is crucial to operate the wind energy
ISSN: 2085-6350

conversions system near the maximum power point to

increase the efficiency of the wind turbine system. Over the
years, control of process and system in industry is
customarily done by experts through the conventional PID
control techniques. This is as a result of its simplicity, low
cost design and robust performance in wide range of
operation conditions [1].
Although the PID controllers have gained widespread
usage across technological industries, it must also be
pointed out that unnecessary mathematical rigorists,
preciseness and accuracy involve with the design of the
controllers have been a major drawback. This has made it
difficult if not impossible for designers, engineers and
technology experts to design intelligent complex system,
non linier systems that can satisfactorily behave as expected
while operating in the human-machine interface [2].
However, various techniques and modification to the
conventional PID controllers have been employed in order
to overcome these difficulties, this include the use of auto
tuning PID controllers, adaptive PID controllers and also
the implementation of compensation schemes to the PID
controllers[2]. Another alternative technique is the use of
unconventional control techniques such as fuzzy logic
control technique will be considered in this study. Hence by
strengthening fuzzy logic controller the searching and
attainment of fuzzy logic rules and high-performance
membership functions will be easier and father [4].
Although the benefits of harnessing the capabilities of FLC
are huge, research efforts on fuzzy logic rules, membership
functions and other parameters are challenging. Research
effort focused majorly on the on the optimal tuning of
membership functions and other parameters were optimized
using FLC. FLC has found application in many scenarios

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

37

like Tracing MPPT, temperature control, robotics and in

many other control systems.
This study employs the fuzzy logic technique to design
a Proportional Integral Derivative (PID) controller and
optimizes the inference rules, membership functions and
scaling of this controller. The performance of the FL
controller is compared with that of the conventional PID
controller. The MATLAB/SIMULINK software forms part
of the modeling and design tools employed in this research.

Expected output voltage is still kept constant, although

some variations occur wind speed. In the cases when the
wind speed is too low, the generator will produce voltage
and power down, the situation so the action is a 'pitch' on
the blade will work to maximize the power output turbine
[1] [6]. To control system load required to converter [7] [8],
and to control the overall system should be developed to the
network and the load [9] [10]. Where is the discussion
involved in the converter control system. Wind turbine
scheme is described as follows.Wind turbine performance is
highly dependent on wind speed average is expressed in
m/s, Then the wind and rotate the blade rotor speed in rpm.
The wind turbine cut-in (from rotating) and cut-out
(discontinued) in a certain wind speed. For example for
speed 3 m/s and the cut-in and 12 m/s cut-off where this
condition is highly related to the speed of wind turbine rotor
speed and generator power. Potential in wind speed is
shown in the Table 1.

Class

Wind
Speed
(m/s)
2.5-4.0

Specific
Power
(W/m)
<75

Capacity
(kW)

Location

0-10

Medium
Sclae

4.0-5.0

75-150

10-100

Big
Scale

>5.0

>150

>100

Java,NTB,
NTT,Maluku,
Sulawesi
NTB,NTT,
Sulsel,Sultra,
South Java.
Sulsel,NTB,
NTT.South
Java.

Small
Scale

*) Source: Lapan

II.

WIND TURBINE MODELING

A. Wind Speed Modeling

Wind model is a description of wind speed
fluctuations, which affect the quality and characteristics of
power control in the wind farm. Fluctuation of wind speed
is power fluctuations affect the resulting wind turbine. Wind
speed patterns can be depicted in the graph to time
measurements. Measurement data that have a relationship
modeling can be searched with several calculations, such as
regression. So that the wind speed can be expressed in the
following equation: V = a + b x + c x2, where V is wind
speed, a and b constant, x is a function of wind speed
against time. Wind speed functions as follows.
B. Wind Turbine Aerodynamic Modeling
The Essential a turbine engine that is a change kinetik
energy from air movement (wind) energy into a first
mechanical axis in the turbine and then into electricity.
Fig.1. is a explain wind energy conversion into mechanical
energy and become to electricity power. Interaction of the
turbine and the wind is very complex, but very cogent may
be simplified with the modeling power or torque
aerodynamics described above. Style is a form of wind
power and aerodynamics to provide lift thrust on the bladeblade. Where is here produce torque on the wind turbine
rotor as :

Wind speed (m/s)

A simple method that run in the wind turbine generator

is used next indukasi connected to the electricity network.
Pembahansan system is so comprehensive and complex. It
is a discussion-only umunya It is like a turbine and
generator, modeling the control of a wind turbine generate
electricity, modeling control trubin wind, the source
disturbance. Maching turbine generator with the generator
[4], MPPT Control [5 ] the all discution to generate
electricity only and does not control the electrical energy to
be raised and used. This research are presented in the study
how the energy raised by the wind to generate electricity
and how this energy is used at the time. Kinetik wind
energy is drive the next turbine transmition generator with
further converted into electrical energy. Generator used in
this system is the "35 V DC motor." energy from here next
conection on the type of converter, the buck converter is
30V<Vin <40 V and the expected output is 12 <Vout <15
V. the output voltage converter and the load on the battery.
The battery used is the type of Lead Acid Dry Cell, 12V
voltage, Capacity 68 Ah.

Fig. 1. Sceme of wind turbine

Table I.
Energy Potential grouping , using and potential location

t ( hour)
Fig. 2. Wind Speed (m/s)

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

38

Proceedings of CITEE, August 4, 2009

Trot = Paero/rot= (1/2) R3 CP (, pitch)

(1)

With P(aero) is power raised on the aerodynamics of the

main axis of the wind turbine that has fingers and have R.eq
wind speed and air density is . Then the wind power can be
shown as follows :
P aero = R2 ueq 3 Cp ( ,pitch )

Ta = J1 t + Bl l + Tg

(3)

T aux = Ta - Bl t

(4)

t = 1/Jt (Taux - Tg )

(5)

Torque aerodynamics in a wind turbine is a function of non

linier, tip speed ratio and the pitch-angle . However, the
point of wind turbine is set, the point of node is constant,
therefore only torque dynamic depends on the tip speed
ratio:
Ta = /2 { R3 vt3 cT () }

(6)

With the method linierisasi around the operating point by

using the first term of the Taylor line, the equation (6) can
be shown :
Ta = vt + t

(7)

= Ta/vt OP=Co vo [2cTO o cT,v ] OP

(8)

=Ta /t OP
= Co vo R cT,w ] OP,Co = R3

(9)

If the Tg and l is an input and output of the system (3),

and (7) then can be arranged into equality (10) and (11) as
follows:
Ta - Tg = ( t J1 + Bl t )

(10)

Ta = vt + t

(11)

The transfer function from the closed plant Generator is:

(12)

C. The Actuator Modeling

The transfer function of the DC motor as actuator
pitch angle is used in the modeling of the PID
controllers. The overall actuator system model is
represented is represented in form[1] ;
Transfer
function
=
2/(S2+12S+24)
(13)
ISSN: 2085-6350

rL
rC

(2)

the air density is dependent on temperature and air

pressure. Cp the power coefficient is undimention
(depending on , pith) which is a representation of the
coefficient of the turbine-rotor. Coefficients of mathematics
for modeling sequence is derived torque aerodynamics,
torque 'auxiliry' comparison and the acceleration as follows:

t (s) / Tg (s) = 1/ ( Jt s Bt + )

D. The Buck Converter design

For power regulated system; The modeling buck
converter is shown in Fig.3-4.[6,7]

Fig. 3. The time in ton at circuit at buck converter

rL
rC

Fig. 4. The time in t off at buck converter

The transition time on, D

Vd = iL rL + L di/dt + vout

(14)

iL

iC + vout / R

(15)

vout = VC + iC rC

(16)

vout = ( iL - iC ) R

(17)

For condition of ttransition signal in buck converter ,

the transition time off , ( D-1);
0 = iL rL + L di/dt + vou
iL

(18)

iC + vout / R

(19)

vout = VC + iC rC

(20)

vout = ( iL - iC ) R

(21)

The value of converter variables are : 15 V < Vin < 20

V( input ) ; 11.9 V<Vout<12.1V(output), voltage set point
charging 12 volt ; Frequency off control signal = 100 k Hz.;
Duty Ratio Limit = 0,35 ( D = .35) ; Capacitor = 10 nF = C
; n = V sat/Vpp = 1,6 ; R = 100 kOhm ; nR = 160
kOhm = rL ; Ri = Tn/4C = 416 ohm. = rC ; L = 0,0001 H.
From equetions (20) & (21) ;
C = C1 D + C2 (1D ) C = 0,35 C1 + 0,65 C2
C2 = [ rC cd
cd ]
=Ax + B u
A = A1 D+ A2 (1-D ) A = 0,35 A1 + 0,65 A2
B = B1 D + B2 ( 1-D ) B = 0,35 B1 + 0,65 B2
Vout = C x
C = C1 D + C2 ( 1 D ) C = 0,35 C1 + 0,65 C2
Vout = C x , C= 0,35 C1+0,65 C2C1= C2, and ;

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

39

transfer function of the DC motor. The error signal of the

system is also monitore using a scope. The overall system
model is representation in form of block diagram in Fig.5.

cd ]
C = [ rC cd
cd =R/(R+rC)=100.000/(100.000+416)=0,996
and C = 0,35 C1 + 0,65 C2
C1 = [ rC cd
cd ] = [ 414
0,996 ]
C2 = [ rC cd
cd ] = [ 414
0,996 ] and,
C = 0,35 C1 + 0,65 C2 = [ 414
0,996 ]

Kontrol
Sudut blade

State space off converter circuit;

0
A1 =

-0,03

C = [ 414

Model
Angin

-43
B1 =

Model
Transmisi

Model
Aero
dinamik

Konv
erter
+
Bater
ai &
Beban

Gene
rator

23,6

0,996 ]
Fig. 5. Structure of a model of variable wind speed
2

Deferential left side row = s X+ 0,03 sX- X

differentiation the righ side ,

and
|u|

x = x1 + co u

0.10
Constant1

co = bo = 1
c1 = b1 - a1 b0= -43 (-0,03)x 1 = -42,07
c2=b2-a2b0-a1(b1-a1b0)=23,6(1x1){(-0,03)
*(43+0,03)}
= 23,611,29=21,3= s2 u-42 s u + 21,3 u

31

Scope2

Subtract1

Add

Slider
Gain1
Slider
Gain2

Transfer Fcn
Slider
Gain

du/dt
Derivative

0.51

1
.1s+.3

X(s)/U(s)=(s2+0,03s -1/ ( s2 -42 s+21,3 )

s+.95

Integrator

the transfer function buck converter :

0.9

1
s

Gain1

Gain2

Display

1
s
Integrator1

Gain
1025
Step

0.04767

Out1

Abs

(22)

Slider
Gain3

Transfer Fcn2

Scope

0.9

and tranfer function plant =

{2/(S2+12S+24}*{(s2+0,03s-1/(s2 -42 s+21,3 )}

Fig.6. PID controller for generator rotor speed

TF=(2s2+0,06s-2)/(S4-30S3-459S2-751S+459) (11)

F. Trasnmision system modeling;

TF = 1/ ( 1-0.51)
III.

0.16

(23)

Amplitude

E. Sensor Modeling
Transfer function of the sensor:[8].
TF = 1/0.95 S + 1

0.18

0.14
0.12
0.1
0.08

(24)

DESIGN OF CONTROL SYSTEM

The MATLAB Simulink software is used in the analysis

of the controller by studying the response generated from
the modeling and simulation of the controller. The simulink
model of the PID controller shown in Fig 5. The difference
between the desired speed and the output feedback is passed
as input into the PID controller subsystem. PID controller
subsystem Fig.5. contains the proportional gain scaling
factor (Kp), the derivative gain scaling factor (Kd) and the
integral gain scaling factor (Ki). The derivative gain factor
and the integral gain factor are both passed through a
derivative bloc and an integral block respectively before
being summed up with the proportional gain factor. The
output of the PID controller subsystem serves as an input to
the motor DC block. The DC motor block is simply the

0.06
0.04
0.02
0

10

Time (s)
Fig. 7.Response of the PID controller for generator
Table II. Performance indicator for conventional PID controller.
Matrices
Rise time
Settling time
Overshoot
Peak
Error Steady state

Conference on Information Technology and Electrical Engineering (CITEE)

Value (s)
0.05
1
0,09
1
0.02

ISSN: 2085-6350

40

Proceedings of CITEE, August 4, 2009

|u|
0.002

Constant1

.91

Gain1

Integrator1
0.9
Add

Slider
Gain1
Slider
Gain2

Transfer Fcn
Slider
Gain

du/dt

Gain2

s+.95

Integrator

595

Display

1
s

76

Step

1
s

Scope2

Gain

Subtract1

0.008861

Out1

Abs

for the fuzzy proportional-derivative controller contains a

set of fuzzy logic rule that define the behavior of the system
in relation between the error signal, error derivative signal
which is the the deference between the desire speed and the
actual speed of the motor. The error derivative signal is
achieved by differentiating the error signal before passing in
to the fuzzy logic controller block. Since the fuzzy logic
controller block. Since the fuzzy logic controller block
expects two inputs, a multiplexer is used to combine the
error signal and the error derivative signals as input into the
block.
The control signal output of the fuzzy logic controller is
passed as the input of the DC motor block to determine the
speed of the motor. The output of the DC motor block
which is the motor speed is monitored using a scope to
examine its response. The desired speed of the motor is set
to a unit constant value to serve a step response input to the
system. The error signal of the system is also monitored
using a scope. The tuning of the fuzzy logic controller can
be achieved by either adjusting the range of the universe of
discourse for the linguistic variables, adjusting the input and
output scaling gains of the controller or adjusting number,
type and positions of the membership function used. The
rule for the fuzzy inference system is shown in Table IV.

Derivative

s2 +.3s+1

.1s+.3

s2 -42s+22

Transfer Fcn2
Slider
Gain3

0.51

Transfer Fcn1

Scope

0.8

Fig.8. Diagram block PID controller with Buck Converter

0.35

Amplituda

0.3

0.25

0.2

0.15

0.1

1.

0.05

10

Scope1

Time (s)

|u|

1
s

Abs

Integrator

Out1

Display

Fig. 9.Response of PID controller with Buck Converter

Table III. Performance for conventional PID controller with converter
Matrices
Value (s)
Rise time
0.4
Settling time
15
Overshoot
0,5
Peak
6.5
Error Steady state
0.04

0.95

1
Constant

z-1

Fuzzy Logic
Controller

Gain

0.5z

0.80

Gain1

z-0.5
Transfer Fcn
First Order

Gain2

Scope

Difference

Fig. 10. Simulink model for Fuzzy Logic Control

Table IV. Fuzzy Logic Rulers

0.44

N
RB
RM
NC

Z
RB
NC
IM

P
NC
IM
IM

The PID controller model is hand tuned by first

increasing the value of the proportional gain Kp, until the
desirable response is obtained. The derivative gains
improves and optimize the response of the system. The
output response of the controller is shown in Fig.6.A fairly
optimal response is achieved for proportional gain value of
.91, integral gain value 76 and a derivative gain value a
595. Fuzzy logic controller; The simulink model of the
fuzzy logic proportional-derivative controller is shown in
Fig.8.
The model consists mainly of fuzzy logic controller
block and the DC motor block. The fuzzy logic controller
block contains a reference to a fuzzy logic inference
system. The inference system has three linguistic variables
which are two inputs (error signal and error derivative) and
the output (control signal). The fuzzy logic inference system
ISSN: 2085-6350

0.42

Amplitude

error/error
N
Z
P

0.4
0.38
0.36
0.34

10

12

14

16

18

20

Fig.10. Output Response for FLC.

Table V. Performance Indicator for FLC
Matrices
Rise time
Settling time
Overshoot
Peak
Error Steady state

Value
0.03
6
0.001
4
0.001

The modified Simulink model of the system showing the value of IAE is
shown in Fig.10.

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

41

overshoot, settling time and rise time are the main

considerations. Matlab where utilized the encoding, testing
and decoding of each of the tuned FLC parameters. The
includes the fuzzy logic rule base, the membership function
definition of the linguistic variables and scaling gain of the
controller was adjusted over a range of 0-250 while the
proportional input scaling gain and the derivative input
scaling gains were adjusted over a range of 0-50 and 0-1.5
respectively. The suitability of the range of scaling gains
was determined from the prior hand tuning of the controller.

1
Scope1

Out1

|u|

1
s

Abs

Integrator

10
Display

0.95

0.80

Gain1

Constant

z-1

Fuzzy Logic
Controller

Gain

Gain2

0.5z
z-0.5
Transfer Fcn
First Order

Scope

NB

NM

-1

-0.6

NS Z

PS

PM

PB

Difference

Fig. 11. Modified simulink model for PD-FLC with the inculcation of IAE

NB

NM

NS

PS

PM

PB

-0.2 0

Input variable dt

0.2

0.6

Fig.13. Modified Membership plot Function for Input variable dT (rate of

change in error)

-1

-0.6

-0.2 0 0.2
Input variable data

0.6

NB

NM

-1

-0.6

NS

PS

PM

PB

Fig.12. Modified Membership plot Function for Input variable delta (error)

( IAE =

e ( t ) dt ) is used as a measure of the system

performance since it known to give a better all round

performance indicator of a control system response where

-0.2 0 0.2

0.6

output variable speed

Fig.14. Modified Membership Function plot for Output variable speed

Table VI. Optimized fuzzy logic rules

0.41
0.4
0.39

Amplitude

Figure11. shows the system output response after

considerable hand-tuning of the fuzzy logic PD-controller.
output
response
after
considerable
hand-tuning.
Optimization of controller: In essence, the beauty of this
study lies predominantly in the use of an effective, eminent
and fast-spreading optimization techniques in the
optimization of the multivariable PD fuzzy logic controller
developed earlier.
The PD-FLC for motor speed is re-implemented
but the only difference this time is that is optimized. The
fuzzy inference system consists of three linguistic variables
(two inputs and one output) each having seven membership
function sets. This results in 49-rule fuzzy inference system
with input as the error and the rate of change in error. The
output of the fuzzy logic inference system is the control
action of the controller and the inverse of all the variables
set within the range (-1,1).The performance indicators
obtained from the time response plot of FLC are shows in
Table 15.
The Fuzzy Logic Controller : The tuning approach employs
the use of MATLAB M and functions to gains, run the
Simulink based simulation, check the resulting performance
and continuously modify the fuzzy inference system for a
number of time in search for an optimal solution Figure
9.The integral of absolute error ;

0.38
0.37
0.36
0.35
0.34

Time (s)

10

Fig. 15. Step response of optimized fuzzy logic PD Controller

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

42

Proceedings of CITEE, August 4, 2009

Table VII. Performance indicator for FLC

Matrices
Rise time
Settling time
Overshoot
Peak
Error Steady state

Value
0.8
1.8
0.001
1
0.001

IV. Resuts And Discussion

The performance metrics of the conventional PID
controller and FLC obtained from the simulation of the
motor speed control and converter are shown in Table 6.A.
The most desirable performance requires the controllers to
have the smallest possible value for the rise time, overshoot
and the settling time. It is also required for the final value
should be close as possible to desired value which is unity.
From the table, it can see that the FLC can produce a
desirable response performance with the use of only PID.
This is contrary to the implementation of conventional
controller which requires the use of proportional, derivative
and integral component before a desirable or satisfactory
response can be obtained. When compared to the
conventional PID controller, the FLC shows a better
performance in term of overshoot while it exhibits a slightly
lesser performance in terms of rise time and settling time.
Table VIII. Comparation of performance indicator for PID Conventional
and fuzzy logic controller and FLC +IAE
Performance
Indicator
Rise time
Settling time
Overshoot
Peak time
Error Steady state

PID
controller
0.05
1
0,09
1
0.02

PID
Controler+
+converter
0.4
15
0,5
6.5
0.04

FLC
0.03
6
0.001
4
0.001

FLC +
IAE
0.8
1.8
0.001
1
0.0001

V. CONCLUTION
This study has succeeded in design of a fuzzy logic
controllers. It was also shown through simulation that the

ISSN: 2085-6350

fuzzy logic controller is performing better than a

conventional PID controller when both controllers are
subjected to the same operating conditions. The performance
metrics taken into conditions are the overshoot, rise time and
state error. MATLAB/SIMULINK is used to simulate and
fine-tune the controller models. The simulation results a
performance of systems show that the FLC is functioning
better than a conventional PID controller in terms of the rise
and settling time aafter the text edit has been completed, the
paper is ready for the template. Duplicate the template file
by using the Save As command, and use the naming
convention prescribed by your conference for the name of
your paper. In this newly created file, highlight all of the
contents and import your prepared text file. You are now
ready to style your paper; use the scroll down window on the
left of the MS Word Formatting toolbar.
REFERENCES
[1].

Ali Musyafa,et.all. Aplication of Steam Turbine frequency control

using Neural Network, Gematek Jurnal Teknik Komputer Volume 9
No.2 September 2007.
[2]. Ali Musyafa, et.all. Local Shot-Term Wind Speed Prediction in the
Nganjuk City (East java) using Neural Network, Prociding
International conference NAE, Lombok, Mei 2009.
[3]. Ali Musyafa, et..all. Kajian Potensi Tenaga Angin di Tiga Lokasi,
Sampang, Mojokerto dan Nganjuk di Jawa Timur, Prociding
National conference , FTI -ITS, Juli 2009.
[4]. Ali Musyafa, Measurement Uncertainty of wind speed condition in
three location in East Java, SITIA-2009
[5]. Kumar,M. And D.P. Garg, 2004. Intelegent learning of FLC via NN
and GA. Proceedings of the Japan-USA symposium on flexible
Automation, July19-21, Colorado,pp:1-8.http://www.duke.edu/
manish /UL_029.pdf.
[6]. Tang, K.S., K.F. Man, G. Chen and S. Kwong, 2001. An optimal
fuzzy PID controller, IEEE Trans.Ind.Elect. 48: 757-765.
DOI:10.1109/41.937407.
[7].
S. Jiao, Bosst Cconverter Design For 20 KW Wind Turbine
Generator, NT Centre For Energy , Darwin NT.0909, 2008
[8]. Lucian Mihet ; Wind Turbine Generator Modeling and Simulation
Where Rotational speed in the Controlled Variable IEEE 2004.
[9] MD Arifujiyaman; Modeling and Control of Small Wind Turbine
CCECE/CCGEL IEEE 2005.
[10] Bongani Malinga;Modeling and Control of a Wind Turbine as a
Distributed Resource IEEE 2003.
[11] B.Neammanee, Control Performance Analysis of Feedforward and
Maximum Peak Power Tracking for Small-and Medium sized
Fixed Pitch Wind Turbine, ICARCV-IEEE 2006.

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

43

Optimal Power Flow Analysis Using Genetic

Algorithm in 500 kV Java Bali Interconnection
System
Buyung Baskoro, Adi Soeprijanto
Electical Engineering Department, Sepuluh Nopember Institute of Technology
Surabaya 60111, Indonesia
[email protected], [email protected]

Ontoseno Penangsang
Electical Engineering Department, Sepuluh Nopember Institute of Technology
Surabaya 60111, Indonesia
[email protected]
AbstractElectricity is generated by power plants then
transmitted through transmission system and distributed to
various electrical loads among electrical power system
network. Since those loads consume electricity, the electricity
is being generated. Optimal power flow analysis is a
computation to minimize objective functions such as
generating cost or transmission loss by controlling active and
reactive power of each interconnected power plant considering
specific constraints. Common used constraint is minimum and
maximum constraints of active and reactive power generation.
This research simulates the calculation of optimal power flow
analysis considering generator capability curve of each power
plant in 500 kV Java Bali interconnection system using genetic
algorithm as computational method of optimization. The
purpose of this research is to obtain optimal generation cost
and line losses in certain loading condition and to improve
optimal power flow analysis constraint using generator
capability curve. The result shows that genetic algorithm is
able to produce lower generation cost and transmission losses
in the system.

realistic constraint of a power plant capability in electrical

generation.

KeywordsOptimal Power Flow

Algorithm, Generator Capability Curve

A. Type of Buses
The type of buses can be divided into three kinds, there
are as follows [6].

I.

Analysis,

Genetic

INTRODUCTION

Electricity is generated by power plants then transmitted

through transmission system and distributed to various
electrical loads among electrical power system network.
Since those loads consume electricity, the electricity is being
generated. In the practical electrical power system, power
plants are not placed in the same distance from loads and
generation cost of each of them is also different.
Optimal power flow analysis is a computation to
minimize objective functions such as generating cost or
transmission loss by controlling active and reactive power of
each interconnected power plant considering specific
constraints. Common used constraints are minimum and
maximum constraints of active and reactive power
generation. Those mentioned constraints do not show the
real characteristic of a power plant. Generator capability
curve is a characteristic curve that is able to show more

Genetic algorithm is a new computational algorithm for

optimization problem inspired by theory of evolution and
then adopted to seek solution of certain problems. This
algorithm is able to give global solution of a computation
that classical optimization method techniques cannot be
applied. One of genetic algorithm applications is
combinational optimization problem that is mainly purposed
to get optimal solution of a problem that has many solution
possibilities.
The purpose of this research is to obtain optimal
generation cost and line losses in certain loading condition
and to improve optimal power flow analysis constraint using
generator capability curve.

II.

SUPPORTING THEORY

Slack bus: Slack bus or swing bus is a bus used as

reference bus in the system. The voltage magnitude
and phase angle is maintained constant. This kind of
bus gives additional power between total load and
generated power caused by line losses.
Load bus: In this bus, active and reactive power is
maintained constant. The value of voltage magnitude
and phase angle is calculated.
Generator bus: In this bus, active power and voltage
magnitude is maintained constant. The phase angle
and reactive power is calculated.
B. Newton Raphson Method
For large scale electrical power system, Newton Raphson
method is proven more efficient and practical in order to
obtain current flows in transmission network and power flow
through buses. The complex form in bus i is stated by using
formula as below [6].

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

44

Proceedings of CITEE, August 4, 2009

Pi jQi = Vi i

New estimation for bus voltage is

V j ij + j

ij

j =1

(1)

Separation in real and imaginary parts yields

n

Pi =

V V
i

Qi =

Yij cos(ij i + j )

(2)

(16)

The iteration process will stop if following equation is

satisfied.
Pi (k )

V j Yij sin( ij i + j )

j =1

(3)
J 2

J 4 V

P J 1
Q = J
3

(15)

Vi ( k +1) = Vi( k ) + Vi( k )

j =1
n

i( k +1) = i( k ) + i( k )

(4)

(17)

Qi(k )

(18)

C. Optimal Power Flow Analysis

1) Lagrange Optimization Method

The element of J1 is stated using

Pi
=
i

V j Yij sin( ij i + j )

F1

j 1

P1

Transmis
idengan
lossesPL

(5)
F2

Pi
= Vi V j Yij sin( ij i + j )
j

P2

ji

PD

(6)

FN

PN

The element of J2 is stated using

Pi
= 2 Vi Y ii cos ii +
Vi

Pi
V j

= Vi Yij cos ij i + j

Yij cos ij i + j

j i

Figure 1. N Units Thermal Power Plant Serves Load PD

(7)

ji
(8)

Mathematically, this problem can be written as an

objective function stated in FT having the same active
power value to the total generation cost supplying load plus
transmission losses [5].
FT = F1 + F2 + F3 + ... + FN

The element of J3 is stated using

Qi
=
i

V j Yij cos( ij i + j )

(9)

ji

(10)

The element of J4 is stated using

Qi
= 2 Vi Y ii sin ii
Vi

Qi
V j

Yij sin ij i + j

j i

= Vi Yij sin ij i + j

ji

(
N

F(Pi ) = min

i =1

+ i Pi + i Pi 2

i =1

Pi(k) = Pi sch Pi(k)

(13)

Q i(k) = Qisch Qi(k)

(14)

(20)

Lagrange equation can be written

L = FT +

(12)

(k)
Pi and Qi(k) is difference value between input value
and calculated value called power residual.

ISSN: 2085-6350

The solution of the problem is how to minimize FT

while considering constraint of total generated power that is
should be in the same value with total load including
transmission losses. The mathematic formula is written as
following
min

(11)

(19)

i =1

j 1

Qi
= Vi Yij cos ij i + j
j

F(P )

(21)

P
P
P
L Fi
+ D + L i
=
Pi Pi
Pi Pi
Pi

=0

(22)

P
L Fi
+ L 1 = 0
=
Pi Pi

Pi

(23)

=0

(24)

P
L Fi
1 L
=
Pi
Pi Pi

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

P
Fi
= 1 L
Pi
Pi

45

(25)

While,

PD is independent to the changing value of Pi , then

PD
=0
Pi

PL changes depending to the current value and

PL dPL

transmission impedance network, then

Pi
dPi
2) Power Balace Constraints

The line impedance or Z in the system is stated in unit of

(Ohm). In order to ease the calculation, then the unit is
converted into p.u (per unit) form. The calculation of Z
value in p.u is given by following formula.
Z
Z p.u = ohm
(29)
Z base
3) Evaluation of System Model
After choosing bus and line parameters, the system is
evaluated using power flow analysis to obtain information
about system model stabilization. The results give the value
of voltage magnitude between 0.95 and 1.05, thus the model
is proven right.

P = P
i

+ PL

(26)

i =1

3) Minimum and Maximum Active Power Constraints

Pi min Pi Pi max
(27)
4) Generator Capability Curve

Figure 2. Example of Generator Capability Curve

D. Genetic Algorithm
Genetic algorithm is a computational algorithm for
optimization problem inspired from theory of evolution to
seek the solution of certain problems. There are lots of
variations in genetic algorithms, one of them is genetic
algorithm for combinational optimization. It is mainly
purposed to get optimal solution value due to problems that
have many solution possibilities. In this case, genetic
algorithm is used to optimize generation cost through power
plant combination in 500 kV Java Bali interconnection
system.
III.

METHODOLOGY

A. System Modelling
1) Bus Parameters
The one line diagram of 500 kV Java Bali
interconnection system is shown in figure 3. The MVA base
is set to 1000 MVA, and kV base value is 500 kV.
2) Line Parameters
The other preset bus parameters are value of Z base. The
Z base value in transmission lines can be curtained using
formula
(kVbase ) 2
Z base =
(28)
MVAbase

Figure 3. Topology of 500 kV Java Bali Interconnection System

4) Cost Function
Formula to obtain generating cost function of each power
plant in 500 kV Java Bali interconnection system is
F = H cos t

(30)

H is characteristic value of heat rate stated in unit Btu/h,

while cost is fuel cost stated in unit Rp/MBtu [5]. After
obtaining generation cost in several points of active power,
then interpolation process in those points of active power
generating cost is done. Thus, exponential equation is
obtained. Those obtained equations are the generating cost
characteristic function of power plants.

B. Lagrange Optimal Power Flow Analysis

One way to include the effects of transmission losses in
power system network is by stating the transmission losses
as a function of quadratic active power output of power plant
[6]. The simple quadratic equation is

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

46

Proceedings of CITEE, August 4, 2009

N

P B

PL =

ij P j

(31)

i =1 j =1

More common formulas consist of linear and constant

equations known as Kron loss formula.
N

PL =

Pi Bij P j +

i =1 j =1

0i Pi

+ B 00

i =1

(32)

The coefficient of B ij is losses coefficient or B

coefficient. This B coefficient is maintained constant.

Furthermore, the optimal power generation can be stated as
follows
N

FT =

Pi =

i =1

i =1

+ i Pi + i Pi 2

(33)

In order to margin the generated power is equal to total

load including transmission losses, thus
N

P = P
i

+ PL

i =1

(34)

The inequality problem can be written as follows

i = 1,...,N
Pi ( min ) Pi Pi ( max )
(35)
Additional information whether Pi ( min ) and
Pi ( max ) are minimum and maximum generated active
power of power plant i.

C. Genetic Algorithm Optimal Power Flow Analysis

1) Genetic Algorithm Parameters
Parameters used are as follows
Genes
:8
Population
: 100
Crossover
: 0.8
Mutation
: 0.3
Maximum Generation
: 50
2) Population
A matrix of population genes with the value of each
element is a random real number between 0 and 1 is
generated. Inside that population, a row means a species,
each species contains chromosomes, and each chromosome
contains several genes.
3) Chromosome Encoding
The used encoding scheme is real number encoding.
Each chromosome in the population is decoded into real
value of generated active power of power plant considering
minimum and maximum value of generated active power.
P = MVAmin + (MVAmax MVAmin ) chromosome
(36)
4) Neural Network
Minimum and maximum value of generated reactive
power in power plant is determined by input value of
generated active power through neural network. After
decoding certain chromosome into real value of active
power, neural network provides the minimum and

ISSN: 2085-6350

maximum reactive power as processed from value of

generated active power.
5) Fitness
Fitness value is determined by total generation cost and
transmission losses. The objective function is to seek
minimum generation cost and transmission losses, thus if all
constraints in optimal power flow analysis are satisfied,
then fitness value is calculated from those variables.
6) Selection
Selection process is done using the system of weighted
roulette wheel of all chromosomes with its specific fitness
value. Chromosomes with high fitness value have high
weight value in roulette wheel however chromosomes with
low fitness value have low weight value. The total
percentage of all fitness value equals to 100%.
7) Crossover
Crossover is carried out to avoid premature convergence
happens. The frequency of crossover operator is controlled
by value of Pc . The higher value of crossover probability,
the faster for new structure is created in the population.
8) Mutation
Mutation is used to do modification for one or more
gene values in the same chromosome. Mutation is done
randomly in genes with a low probability.
IV.

SIMULATION RESULTS

A. System Evaluation
Before the program is implemented in the real system of
500 kV Java Bali interconnection system, it is tested by
using comparison between Lagrange method and genetic
algorithm in IEEE 30 bus system. The results of Lagrange
method is given in table 1, while the results of genetic
algorithm is shown in table 2. Total load supplied in this
IEEE 30 bus system is 189.2 MW.
TABLE I.

P1
P2
P3
P4
P5
P6
Total

TABLE II.

LAGRANGE METHOD IN IEEE 30 BUS SYSTEM

Lagrange
MW
MVAR
44.66
-5.34
58.84
36.38
23.32
11.94
29.12
36.08
17.55
9.28
18.43
12.45
191.91
100.78

2.71

Cost
$/h
129.20
163.55
57.32
101.70
60.33
63.79
575.88

GENETIC ALGORITHM METHOD IN IEEE 30 BUS SYSTEM

GA

P1
P2
P3
P4
P5
P6
Total

Losses
MW

MW
42.79
57.00
23.00
35.00
17.00
17.00
191.79

MVAR
-5.01
36.35
12.16
34.07
9.70
12.95
100.22

Losses
MW

2.59

Cost
$/h
122.19
156.61
56.06
123.97
58.23
58.23
575.27

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

47

B. Java Bali 500 kV Interconnection System

1) First Loading Condition
Power generation in 500 kV Java Bali interconnection
system is shown in table 3, while the simulation using
genetic algorithm is shown in table 4, and the simulation
using neural network-genetic algorithm is shown in table 5.
TABLE III.

Suralaya
Muaratawar
Cirata
Saguling
Tanjungjati
Gresik
Paiton
Grati
Total
TABLE IV.

REAL SYSTEM POWER GENERATION IN 500 KV JAVA BALI

INTERCONNECTION SYSTEM
Real System
MW
MVAR
3,199.77
1,144.17
1,178.00
1,471.34
629.00
-109.76
634.00
419.33
668.00
425.93
821.00
619.40
2,806.00
1,175.60
0.00
0.00
9,935.77
5,146.01

MW
3,188,93
1,250.64
940.26
691.14
1,179.11
269.69
2,397.57
0.00
9,917.34

TABLE V.

Suralaya
Muaratawar
Cirata
Saguling
Tanjungjati
Gresik
Paiton
Grati
Total

99.77

Cost
Rp/h
590,949,741.48
3,985,914,408.15
3,774,000.00
3,488,268.00
120,324,230.38
740,722,429.28
524,007,928.24
86,557,397.40
6,055,738,402.93

GENETIC ALGORITHM METHOD IN 500 KV JAVA BALI

INTERCONNECTION SYSTEM
GA

Suralaya
Muaratawar
Cirata
Saguling
Tanjungjati
Gresik
Paiton
Grati
Total

Losses
MW

MVAR
1.132,28
1,455.68
-142.51
369.60
390.88
634.95
1,112.63
0.00
4,953.50

Losses
MW

81.34

Cost
Rp/h
591,227,347.05
4,287,817,646.29
5,641,546.06
3,802,668.82
200,454,457.16
232,806,858.34
397,825,268.06
86,557,397.40
5,806,133.189.18

NN-GA METHOD IN 500 KV JAVA BALI

INTERCONNECTION SYSTEM

NN-GA
MW
MVAR
3,593,56
1.152,24
1,113.31
1,480.93
998.51
-148.52
764.69
343.27
739.85
325.43
334.95
751.15
2,368.33
999.41
0.00
0.00
9,913.20
4,903.92

Losses
MW

77.20

Cost
Rp/h
570,360,581.90
3,723,171.661.20
5,991,061.09
4,207,316.39
132,278,875.83
288,744,673.55
389,460,737.44
86,557,397.40
5,200,772,304.82

From the results above, it can be said that in the same

loading condition of 9,836 MW, genetic algorithm can
reduce total generation cost up to 249,605,213.75 Rp/h or
4.12%, meanwhile the neural network-genetic algorithm can
reduce total generation cost up to 894,966,098.11 Rp/h or
14.12%.

2) Second Loading Condition

Power generation in 500 kV Java Bali interconnection
system is shown in table 6, while the simulation using
genetic algorithm is shown in table 7, and the simulation
using neural network-genetic algorithm is shown in table 8.

TABLE VI.

Suralaya
Muaratawar
Cirata
Saguling
Tanjungjati
Gresik
Paiton
Grati
Total
TABLE VII.

Suralaya
Muaratawar
Cirata
Saguling
Tanjungjati
Gresik
Paiton
Grati
Total

REAL SYSTEM POWER GENERATION IN 500 KV JAVA BALI

INTERCONNECTION SYSTEM
Real System
MW
MVAR
3,332.18
988.56
1,470.00
679.36
400.00
484.32
535.00
1,043.09
830.00
361.87
810.00
608.62
2,820.00
895.04
198.00
395.97
10,395.18 5,456.83

113.18

Cost
Rp/h
586,309,219.13
5,243,786.025.07
2,400,000.00
2,943,570.00
146,959,544.62
729,802,889.66
528,641,810.78
504,471,401.82
7,745,314,461.08

GENETIC ALGORITHM METHOD IN 500 KV JAVA BALI

INTERCONNECTION SYSTEM
GA
MW
3,122.15
1,483.08
982.07
695.09
1,087.22
295.33
2,440.80
260.12
10,365.86

TABLE VIII.

Suralaya
Muaratawar
Cirata
Saguling
Tanjungjati
Gresik
Paiton
Grati
Total

Losses
MW

MVAR
1,160.60
1,471.67
-124.53
422.45
358.61
646.34
857.69
378.53
5,171.36

Losses
MW

83.81

Cost
Rp/h
592,595,331.54
5,302,883,057.72
5,892,432.33
3,824,097.09
186,891,580.60
254,647,468.53
410,358,414.37
644,202,156.70
7,401,294.538.87

NN-GA METHOD IN 500 KV JAVA BALI

INTERCONNECTION SYSTEM

NN-GA
MW
MVAR
3,843.33
1,203.56
1,165.63
1,526.62
1,052.01
-131.43
732.56
404.30
1,125.25
337.39
241.21
636.07
1,997.36
838.68
211.81
362.16
10,369.16
5,177.35

Losses
MW

87.16

Cost
Rp/h
546,702,181.19
3,935,226,489.65
6,312,079.21
4,030,518.43
192,548,601.53
208,751,673.92
291,073,482.38
535,182,724.59
5,719,827,750.89

From the results above, it can be said that in the same

loading condition of 10,282 MW, genetic algorithm can
reduce total generation cost up to 344,019,922.21 Rp/h or
4.44%, meanwhile the neural network-genetic algorithm can
reduce total generation cost up to 2,025,486,710.19 Rp/jam
or 26.15%.
V.

CONCLUSION

The evaluation test shows that genetic algorithm method

is able to be one kind of solution in optimal power flow
analysis and produce lower total generation cost and
transmission losses. In the first loading condition, genetic
algorithm reduces total generation cost up to 249,605,213.75
Rp/h or 4.12%, meanwhile neural network-genetic algorithm
reduces total generation cost up to 894,966,098.11 Rp/h or
14.12%. In the second loading condition, genetic algorithm
reduces total generation cost up to 344,019,922.21 Rp/h or
4.44%, meanwhile neural network-genetic algorithm reduces
total generation cost up to 2,025,486,710.19 Rp/h or 26.15%.

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

48

Proceedings of CITEE, August 4, 2009

REFERENCES
[1]

[2]

[3]

[4]

[5]
[6]
[7]

H.W. Dommel, William.F. Tinney, Optimal Power Flow Solutions,

IEEE Transactions on Power Apparatus and Systems, Vol. 87,
10:1866-1876, October 1968.
T. Bouktir, L. Slimani, M. Belkacemi, A Genetic Algorithm for
Solving The Optimal Power Flow Problem, Leonardo Journal of
Sciences, 4:44-58, January-June 2004.
M. Younes, M. Rahli, L.A. Koridak, Economic Power Dispatch
Using Evolutionary Algorithm, Journal of Electrical Engineering,
Vol. 57, 4:211-217, 2006.
R. Ouiddir, M. Rahli, L.A. Koridak, Economic Dispatch using a
Genetic Algorithm: Application to Western Algerias Electrical
Power Network, Journal of Information Science and Engineering,
Vol. 21, 659-668, 2005.
A.J. Wood, B.F. Woolendberg, Power Generation, Operation, and
Control, John Wiley & Sons Inc., Canada, 1996.
H. Saadat, Power System Analysis, McGraw-Hill Book Co.,
Singapore, 1999
R.L. Haupt., S.E. Haupt, Practical Genetic Algorithm, John Wiley &
Sons, Canada, 1998.

Buyung Baskoro, was born in Tuban,

Desember 15th 1987 of Ir. Eko Pratomo
and Ir. Susilorini couple as the first of
two children. He starts his study from
Sidomulyo 2 State Elementary School of
Tuban, then continues to No. 1 State
Junior High School of Tuban, and then
No. 1 State Senior High School of
Tuban and graduates in 2005. Recently,
he is a student of Electrical Engineering
Department of ITS Surabaya majoring in
Power System Engineering and active as an assistant of
Power System Simulation Laboratory.

power system.

ISSN: 2085-6350

Adi Soeprijanto is a doctor and

lecturer in Electrical Engineering
Department of ITS Surabaya. He
finished his bachelor degree in 1988
and master degree in 1995 in
Bandung Institute of Technology. He
achieved his doctor degree in
Hiroshima University in 2001. His
research involves power system
monitoring, power system operation
and control, especially large scale of

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

49

Transient Stability Assessment of Java Bali 500 kV

Multi Machine Electrical Power System Using
Committee Neural Network
Eko Prasetyo, Boy Sandra
Department of electrical engineering, Sepuluh Nopember Institute of Technology
Surabaya 60111, Indonesia
e-mail : [email protected]

Adi Soeprijanto
Department of electrical engineering, Sepuluh Nopember Institute of Technology
Surabaya 60111, Indonesia
e-mail : [email protected]
Abstract In a high scale multi machine electrical power
system, transient stability control of the system is a highly
important thing. Several transient stability assessment
methods needs a long computing time, to overcome it then
people develop transient stability assessment methods using
artificial neural network. In this research, stability status of
the system and the critical clearing time (CCT) of the circuit
breaker (CB) will be predicted using committee neural
network (CNN). This method gives its predicting process into
several individual neural networks which will generate final
decision of the network.
The examined multi machine system in this research is
Java - Bali 500kV interconnection system which is belongs to
PLN. Assumed that there were three phase symmetrical fault
located on the system and we calculate its transient stability
status and CCT using equal area criteria method. Obtained
data from that computation will be used as input and target
for neural network. Each expert will be trained using
backpropagation algorithm and the result will be applied to
assess transient stability condition of the system. To verify the
effectiveness of the proposed method, the result of assessment
using CNN will be compared with the result from computation
using equal area criteria method. Results show that CNN can
achieve more accurate assessment for transient stability status
than CCT.
Keywordstransient stability, multi machine system,
committee neural network, extended equal area criteria,
backpropagation algorithm.

I.

INTRODUCTION

Quality of electrical power system must be maintained

to fulfil its services continuity standard. One of the most
influencing factors to services continuity is transient
stability of the system. Transient stability is the ability of an
electrical power system to reach new stable operating
condition after experiencing of big disturbance [1]. Large
scale blackouts will give very big loss materially or
immaterially [2]. In this research, committee neural network
(CNN) will be used to get the accurate assessment and less
computing time for transient stability assessment of a multi
machine power system.
One type of CNN ever used to assess transient stability
status of a multi machine electrical power system when one

of the transmission lines cut off on a certain fault clearing

time (FCT) [3]. Basically, CNN constructed from a several
neural network experts which is work simultaneously to get
an overall decision [4]. Learning process of each neural
network expert which is constructing CNN will use
backpropagation neural network (BPNN) method. BPNN
use supervised learning method which is use a pair of input
and output to determine the optimum weight and also
designed for feed forward multi layer neural network [5].
Multilayer perceptron (MLP) which has trained using
BPNN was used to assess critical clearing time (CCT) of a
multi machine electrical power system [6]. Beside that,
radial basis function neural network (RBFNN) and
probabilistic neural network (PNN) also used to assess
transient stability status of a multi machine electrical power
system when one of the transmission lines cut off on a
certain fault clearing time and it was proven that it can
assess better than MLP [7].
II.

SUPPORTING THEORY

A. Basic of Equal Area Criteria

Transient stability study involves act of determining
whether the system will remains stable or keep its
synchronous operating point after experiencing of big
disturbance [8]. Big disturbances can be determined as three
phase symmetrical fault, large scale load shedding, or
loosing one of the biggest generators. Loosing
synchronization due to transient fault will be visible after 2
3 second after the first fault [9]. This equal area criteria
(EAC) method based on graphical interpretation of the
power on rotating object as helpful effort to determine
machines capability to maintain its stability after
experiencing of disturbance [10].

Figure 1. Single machine system connected to an infinite bus, three phase

symmetrical fault occurs at point F.

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

50

Proceedings of CITEE, August 4, 2009

Figure 3. Illustration model of a neuron

Figure 2. Equal area criteria curve for three phase symmetrical fault which
is far from sending point.

Assume there is a synchronous generator connected to

an infinite bus and three phase symmetrical fault occurs at
point F which is far from sending point, equivalent
reactance will be increased and reduce power transfer
ability of the system. When the fault occurs, operating point
of the machine will be shifted to point b at curve B.
Mechanical input power which is exceed electrical output
power will make rotor run faster and store kinetic energy
that increase angle . If we assume that fault has already
cleared on 1 by cutting transmission line, it will shift
operating point to point e at curve C. Pure power will be
reduced and also kinetic energy on the rotor will be reduced
until 0 at point f. Using equal area criteria, system will be
stable if area A1 equals to area A2. Mathematically, we can
calculate each area using these equations:

A1 = Pm ( c 0 ) c P2 max sin d
0

A2 = max P3 max sin d Pm ( max c )

c

(1)
(2)

Damping factor on the machine will make oscillating

power disappear and new steady state operating condition
will be started at point f. In a multi machine electrical power
system, equation for equal area criteria can be written using
same rule as a single machine connected to an infinite bus
using several simplification assumptions [10].
B. Basic of Artificial Neural Network
Artificial neural network (ANN) is an information
processing system which has its own characteristic that is
made imitating a real biological neural network [11].
Assume that there are n input and weight, we can determine
output function for each neuron mathematically using this
equation:
F(x,W) = f(w1x1 + ... + wnxn)

(3)

Learning process for ANN is a process to arrange all

value of its weights to get the best or optimal value by
training it using several data pattern according to desired
ability for the network systems [12].

ISSN: 2085-6350

C. Backpropagation Neural Network

BPNN learning algorithm consists of two main
processes, there are feed forward and backpropagation of its
error [5]. Basically, this algorithm is called as
backpropagation neural network because when we give it
input pattern as a learning pattern, this pattern will be
propagated forward to its hidden layer then will be
continued to its output layer which will give response called
networks output. In case this networks output is not equal
with desired output, then this value will be propagated
backward to its hidden layer and continued to its input units.
D. Committee Neural Network
A complex computational task is solved by dividing it
into a number of computationally simple tasks and then
combining the solutions to those tasks [4]. In supervised
learning, computational simplicity is achieved by
distributing the learning task among a number of experts.
The combination of experts is said to constitute a committee
machine or commonly called committee neural network
(CNN). Basically, it fuses knowledge acquired by experts to
arrive at an overall decision that is supposedly superior to
that attainable by anyone of them acting alone.
CNN is a universal approximator, it could be classified
into two major categories, static structure and dynamic
structure [4]. Static structure includes ensemble averaging
where the output of different predictors are linearly
combined to produce an overall output and boosting where
a weak learning algorithm is converted into one that
achieves arbitrarily high accuracy. In the other hand,
dynamic structure includes mixture of experts, in which the
individual responses of the experts are non linearly
combined by means of single gating network and
hierarchical mixture of expert, in which the individual
responses of the experts are non linearly combined by
means of several gating networks arranged in a hierarchical
fashion.
III.

METHODOLOGY

A. Research Plan
Research begins by gathering characteristic data of the
line, bus, and power generation of Java Bali 500 kV
interconnection power system. Furthermore, transient
stability status and CCT of the system will be calculated
using equal area criteria method. These collected data will
be used as training pair for each neural network experts.

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

Figure 4. Flowchart for research plan

Learning process for each expert will be held

individually using backpropagation algorithm. After all
learning process already done, its weights and normalization
parameters will be saved. Furthermore, all saved weights
and parameters will be used as supporting data for portable
application called Intelligent Transient Stability Analyst
which is made using Delphi.
B. Transient Stability Assessment Using Equal Area
Criteria
In case there are occurred transient disturbance on a
large scale electrical power system, commonly there are
several or maybe only one generator which is most
disturbed or affected by that disturbance [10].
Disturbed generators could have different stability
margin, but generator which has the lowest stability margin
can be assumed as the most critical generator. When the
disturbance occurs the most critical generator will be the
first generator that will lose its synchronism.
C. Transient Stability Assessment Using Committee Neural
Network
Active and reactive power generated by each generators,
faulted buses, and also transmission lines which will be cut
off will be used as input for expert 1, 2, and 3. Transient
stability condition of the system which is formed as
transient stability status and CCT will be used as output for
each neural network expert. Composite of an output from
expert 1, 2, and 3 will be used as an input for expert 4 and
overall output of the network is an output of expert 4.

51

Transient stability condition which is formed as stability

status and CCT will be depicted using certain rule. Stability
status of the system depicted using binary value, value 1
means that the system is stable and value 0 means that the
system is unstable. In a stable system, if the system will
always stable at all CCT condition, CCT will be assumed as
1 second. If the system becomes unstable at a certain CCT
value, that value will be used as an output for the network.
In an unstable system, CCT will be assumed as 0 second.
For learning necessity, neural networks parameter must
be given as follows:
1. Learning function : traingdm
2. Layer amount
:
Input layer
: 18 neurons (expert 1, 2, and 3),
6 neurons (expert 4)
Hidden layer
: 2 layers, 49 and 45 neurons
Output layer
: 2 neurons
3. Activation function :
1st Hidden layer
: tansig (bipolar sigmoid
function)
2nd Hidden layer :
logsig
(binary
sigmoid
function)
Output layer
: purelin (identity function)
4. Maximum epoch : 500000 epoch (expert 1, 2, 3, and 4)
5. Minimum error : 1e-9
6. Learning rate
: 0.55
7. Momentum
: 0.77
8. Learning result will be presented after : 5000 epoch
IV.

TESTING AND ANALYSIS

A. Configuration of Java - Bali 500 kV Interconnection

System
Electrical power system which is used in this research is
Java - Bali 500 kV interconnection system which is belongs
to PLN and consists of 8 generators, 23 buses, and 28
transmission lines with the following single line diagram:

Figure 5. Block diagram of CNN which is used on the research

Figure 6. Single line diagram of Java - Bali 500 kV interconnection system.

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

52

Proceedings of CITEE, August 4, 2009

All parameters of the system which is used in this

research are presented at table I and table II.
TABLE I. Parameter of all transmission lines at Java - Bali 500 kV
interconnection system
No
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

Origin Destination
R
Bus
Bus
1
2
0,0006265
1
4
0,0065133
2
5
0,0131333
3
4
0,0015132
4
5
0,0012464
4
18
0,0006942
5
7
0,0044419
5
8
0,0062116
5
11
0,0041114
6
7
0,0019736
6
8
0,0056256
8
9
0,0028221
9
10
0,0027400
10
11
0,0014747
11
12
0,0019578
12
13
0,0069910
13
14
0,0134780
14
15
0,0135339
14
16
0,0157986
14
20
0,0090361
15
16
0,0375396
16
17
0,0013947
16
23
0,0039864
18
19
0,0140560
19
20
0,0153110
20
21
0,0102910
21
22
0,0102910
22
23
0,0044358

/2Bc

Tap

0,0000000
0,0059898
0,0035306
0,0000000
0,0000000
0,0000000
0,0000000
0,0000000
0,0044210
0,0000000
0,0000000
0,0000000
0,0000000
0,0000000
0,0000000
0,0064291
0,0123948
0,0036383
0,0036322
0,0000000
0,0086307
0,0000000
0,0000000
0,0151144
0,0164639
0,0110659
0,0110659
0,0047698

1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1

X
0,0070088
0,0625763
0,1469258
0,0169283
0,0119750
0,0066693
0,0426754
0,0596780
0,0459950
0,0189618
0,0540480
0,0271130
0,0263242
0,0141685
0,0219024
0,0671659
0,1294900
0,1514074
0,1517848
0,0868146
0,3606623
0,0133994
0,0445967
0,1572480
0,1712880
0,1151280
0,1151280
0,0496247

TABLE II. Reactance and reactive power limit for all generators at Java Bali 500 kV interconnection system
Generator

Bus

Ra

Xd'

Suralaya
Muaratawar
Cirata
Saguling
Tanjung Jati
Gresik
Paiton
Grati

1
8
10
11
15
17
22
23

0
0
0
0
0
0
0
0

0,1418445
0,0498565
0,0982143
0,2820233
0,4842950
0,1869112
0,1567810
0,2630799

5,19
1,82
2,86
1,64
3,20
2,54
4,42
2,76

Q (MVar)
Minimum Maximum
-600
2040
-700
1540
-488
488
-140
440
-240
720
-610
660
-840
1920
-302
566

B. Learning of Artificial Neural Network

Learning process for each neural network expert will be
held using 33 generation pattern which is consist of 24 real
generation pattern and 9 self made pattern. There are 54
combinations for fault locations and transmission line cut
patterns for each generation pattern. Totally, CNN will be
trained using 1782 learning patterns.
TABLE III. Minimum MSE for each expert
Neural Network
Expert 1
Expert 2
Expert 3
Expert 4

MSE
0,000177322
0,001123790
0,003316020
0,000063666

All neural network experts were trained until maximum

epoch condition reached because none of them can reach
desired minimum MSE limit.
C. Testing of Artificial Neural Network
Test is carried out directly on the Windows based
application have been made. For example, assume that there
ISSN: 2085-6350

are three phase symmetrical faults occurs on March 30th,

2009 at 7.00 pm when the system experiencing daily peak
load. Fault occurs at bus 1 and transmission line between bus
1 and 2 will be cut off. Views of the application after
transient stability assessment function already executed are
depicted below:

Figure 7. Views of the application after transient stability assessment

function for the system already executed.

1) Testing Case 1
Test is carried out by assuming that the fault occurs at
bus 1 (Suralaya) which is commonly produce highest
power generation for the system. In this case,
transmission line between bus 1 (Suralaya) and 2
(Cilegon) is cut off.
a) Test Using Trained Pattern
Test is carried out using generation pattern that is
already trained to CNN. Testing condition for this
case is generation condition of the system for each
hour at March 30th, 2009.
Transient Stability Status Assessment of The
System
Comparison of transient stability status for an
output of EAC and CNN method for case 1 (a) is
presented at table IV.
TABLE IV. Comparison of stability status for case 1 (a)
No
1
2
3
4
5
6
7
8
9
10
11
12

EAC
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000

CNN Error (%) No

1,000
0,00 13
1,000
0,00 14
1,000
0,00 15
1,000
0,00 16
1,000
0,00 17
1,000
0,00 18
1,000
0,00 19
1,000
0,00 20
1,000
0,00 21
1,000
0,00 22
1,000
0,00 23
1,000
0,00 24

EAC
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000

CNN Error (%)

1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00

Using data of error presented in table IV, we

can conclude minimum and maximum accuracy
value of CNN method when it used to assess
transient stability status for case 1 (a) as follow:
- Minimum accuracy : 100,00 %
- Maximum accuracy : 100,00 %
Graphically, accuracy of CNN in assessing
transient stability status for case 1 (a) can be seen
at figure 8.

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

53

Comparison of transient stability status for an

output of EAC and CNN method for case 1 (b) is
presented at table VI.

1.500

Status Stabilitas

1.250
1.000

TABLE VI. Comparison chart of stability status for case 1 (b)

0.750
0.500
0.250
0.000
0

10

15

20

25

Data keCNN

EAC

Figure 8. Comparison chart of stability status for case 1 (a)

CCT Assessment of The System

Comparison of CCT for an output of EAC
and CNN method for testing case 1 (a) is
presented at table V.

No
1
2
3
4
5
6
7
8
9
10
11
12

EAC (s) CNN (s) Error (%)

0,146
0,144
1,37
0,146
0,146
0,00
0,142
0,136
4,23
0,134
0,131
2,24
0,135
0,136
0,74
0,140
0,140
0,00
0,147
0,147
0,00
0,136
0,135
0,74
0,133
0,130
2,26
0,136
0,125
8,09
0,164
0,146
10,98
0,186
0,183
1,61

CNN Error (%)

1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00

1.250

Using data of error presented in table V, we

can conclude minimum and maximum accuracy
value of CNN method when it used to assess
CCT value for case 1 (a) as follow:
- Minimum accuracy : 89,02 %
- Maximum accuracy : 100,00 %
Graphically, accuracy of CNN in assessing
CCT for case 1 (a) can be seen at figure 9.

1.000
0.750
0.500
0.250
0.000
0

10

15

20

25

Data ke-

CNN

EAC

Figure 10. Comparison chart of stability status for case 1 (b)

CCT Assessment of The System

Comparison of CCT for an output of EAC and
CNN method for testing case 1 (b) is presented at
table VII.

0.350
0.300
0.250
CCT

EAC
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000

1.500

Status Stabilitas

EAC (s) CNN (s) Error (%) No

0,203
0,203
0,00 13
0,214
0,216
0,93 14
0,214
0,212
0,93 15
0,216
0,210
2,78 16
0,194
0,194
0,00 17
0,194
0,191
1,55 18
0,216
0,213
1,39 19
0,198
0,202
2,02 20
0,155
0,151
2,58 21
0,148
0,151
2,03 22
0,159
0,158
0,63 23
0,137
0,142
3,65 24

CNN Error (%) No

1,000
0,00 13
0,979
2,10 14
1,000
0,00 15
1,000
0,00 16
1,000
0,00 17
1,000
0,00 18
1,000
0,00 19
1,000
0,00 20
1,000
0,00 21
1,000
0,00 22
1,000
0,00 23
1,000
0,00 24

Using data of error presented in table VI, we

can conclude minimum and maximum accuracy
value of CNN method when it used to assess
transient stability status for case 1 (b) as follow:
- Minimum accuracy : 97,90 %
- Maximum accuracy : 100,00 %
Graphically, accuracy of CNN in assessing
transient stability status for case 1 (b) can be seen
at figure 10.

TABLE V. Comparison of CCT for case 1 (a)

No
1
2
3
4
5
6
7
8
9
10
11
12

EAC
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000

0.200

TABLE VII. Comparison of stability status for case 1 (b)

0.150
0.100
0.050
0.000
0

10

15

20

25

Data keCNN

EAC

Figure 9. Comparison chart of CCT for case 1 (a)

b) Test Using Untrained Pattern

Test carried out using generation pattern that is not
already trained to CNN. Testing condition for this
case is generation condition of the system at March
2009 for each hour which has taken randomly.
Transient Stability Status Assessment of The
System

No
1
2
3
4
5
6
7
8
9
10
11
12

Conference on Information Technology and Electrical Engineering (CITEE)

EAC (s) CNN (s) Error (%) No

0,215
0,219
1,86 13
0,197
0,150
23,86 14
0,160
0,160
0,00 15
0,155
0,162
4,52 16
0,141
0,144
2,13 17
0,146
0,154
5,48 18
0,140
0,127
9,29 19
0,137
0,152
10,95 20
0,132
0,134
1,52 21
0,177
0,166
6,21 22
0,145
0,121
16,55 23
0,131
0,137
4,58 24

EAC (s) CNN (s) Error (%)

0,216
0,155
28,24
0,213
0,150
29,58
0,220
0,164
25,45
0,155
0,150
3,23
0,158
0,161
1,90
0,147
0,129
12,24
0,137
0,174
27,01
0,147
0,177
20,41
0,142
0,103
27,46
0,153
0,145
5,23
0,157
0,141
10,19
0,143
0,143
0,00

Using data of error presented in table VII, we

can conclude minimum and maximum accuracy
value of CNN method when it used to assess
CCT value for case 1 (b) as follow:
- Minimum accuracy : 70,42 %
ISSN: 2085-6350

54

Proceedings of CITEE, August 4, 2009

Maximum accuracy : 100,00 %

Graphically, accuracy of CNN in assessing
CCT for case 1 (b) can be seen at figure 11.

1.500

Status Stabilitas

1.250

0.350
0.300
0.250

1.000
0.750
0.500

CCT

0.250
0.200

0.000

0.150

10

0.100

15

20

25

Data ke-

0.050

CNN

EAC

0.000
5

10

15

20

25

Data keCNN

EAC

Figure 11. Comparison chart of CCT for case 1 (b)

2) Testing Case 2
Test is carried out by assuming that the fault occurs at
bus 3 (Cirata) which is commonly produce lowest power
generation for the system. In this case, transmission line
between bus 3 (Cirata) and 4 (Gandul) is cut off.
c) Test Using Trained Pattern
Test is carried out using generation pattern that is
already trained to CNN. Testing condition for this
case is generation condition of the system for each
hour at March 30th, 2009.
Transient Stability Status Assessment of The
System
Comparison of transient stability status for an
output of EAC and CNN method for case 2 (a) is
presented at table VIII.
TABLE VIII. Comparison of stability status for case 2 (a)
No
1
2
3
4
5
6
7
8
9
10
11
12

EAC
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000

CNN Error (%) No

1,000
0,00 13
1,000
0,00 14
1,000
0,00 15
1,000
0,00 16
1,000
0,00 17
1,000
0,00 18
1,000
0,00 19
1,000
0,00 20
1,000
0,00 21
1,000
0,00 22
1,000
0,00 23
1,000
0,00 24

EAC
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000

CNN Error (%)

1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00
1,000
0,00

Using data of error presented in table VIII, we

can conclude minimum and maximum accuracy
value of CNN method when it used to assess
transient stability status for case 2 (a) as follow:
- Minimum accuracy : 100,00 %
- Maximum accuracy : 100,00 %
Graphically, accuracy of CNN in assessing
transient stability status for case 2 (a) can be seen
at figure 12.

ISSN: 2085-6350

Figure 12. Comparison chart of stability status for case 2 (a)

CCT Assessment of The System

Comparison of CCT for an output of EAC
and CNN method for testing case 2 (a) is
presented at table IX.
TABLE IX. Comparison of CCT for case 2 (a)
No
1
2
3
4
5
6
7
8
9
10
11
12

EAC (s) CNN (s) Error (%) No

0.461
0.463
0.43 13
1.000
1.000
0.00 14
1.000
1.000
0.00 15
1.000
1.000
0.00 16
0.431
0.434
0.70 17
0.433
0.431
0.46 18
1.000
1.000
0.00 19
0.460
0.457
0.65 20
0.336
0.336
0.00 21
0.320
0.316
1.25 22
0.349
0.355
1.72 23
0.295
0.282
4.41 24

EAC (s) CNN (s) Error (%)

0.318
0.314
1.26
0.318
0.309
2.83
0.307
0.306
0.33
0.290
0.289
0.34
0.290
0.279
3.79
0.305
0.299
1.97
0.324
0.316
2.47
0.298
0.297
0.34
0.289
0.291
0.69
0.293
0.279
4.78
0.358
0.358
0.00
0.420
0.403
4.05

Using data of error presented in table IX, we

can conclude minimum and maximum accuracy
value of CNN method when it used to assess
CCT value for case 2 (a) as follow:
- Minimum accuracy : 95,22 %
- Maximum accuracy : 100,00 %
Graphically, accuracy of CNN in assessing
CCT for case 2 (a) can be seen at figure 13.
1.050
0.950
0.850
0.750
CCT

0.650
0.550
0.450
0.350
0.250
0

10

15

20

25

Data keCNN

EAC

Figure 13. Comparison chart of CCT for case 2 (a)

d) Test Using Untrained Pattern

Test carried out using generation pattern that is not
already trained to CNN. Testing condition for this
case is generation condition of the system at March
2009 for each hour which has taken randomly.
Transient Stability Status Assessment of The
System

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

55

Comparison of transient stability status for an

output of EAC and CNN method for case 2 (b) is
presented at table X.

Maximum accuracy : 100,00 %

Graphically, accuracy of CNN in assessing
CCT for case 2 (b) can be seen at figure 15.

TABLE X. Comparison chart of stability status for case 2 (b)

EAC
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000

CNN Error (%) No

1.000
0.00 13
1.010
1.00 14
1.000
0.00 15
1.000
0.00 16
1.000
0.00 17
1.000
0.00 18
1.000
0.00 19
1.000
0.00 20
1.000
0.00 21
1.000
0.00 22
1.000
0.00 23
1.000
0.00 24

EAC
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000

Using data of error presented in table X, we

can conclude minimum and maximum accuracy
value of CNN method when it used to assess
transient stability status for case 2 (b) as follow:
- Minimum accuracy : 99,00 %
- Maximum accuracy : 100,00 %
Graphically, accuracy of CNN in assessing
transient stability status for case 2 (b) can be seen
at figure 14.
1.500

Status Stabilitas

1.250
1.000
0.750
0.500
0.250
0.000
0

10

15

20

25

Data ke-

CNN

EAC

Figure 14. Comparison chart of stability status for case 2 (b)

CCT Assessment of The System

Comparison of CCT for an output of EAC and
CNN method for testing case 2 (b) is presented at
table XI.
TABLE XI. Comparison of stability status for case 2 (b)
No
1
2
3
4
5
6
7
8
9
10
11
12

EAC (s) CNN (s) Error (%) No

1.000
0.999
0.10 13
0.449
0.466
3.79 14
0.283
0.333
17.67 15
0.337
0.359
6.53 16
0.306
0.280
8.50 17
0.318
0.302
5.03 18
0.302
0.297
1.66 19
0.294
0.344
17.01 20
0.283
0.277
2.12 21
0.392
0.392
0.00 22
0.316
0.289
8.54 23
0.283
0.271
4.24 24

1.050

CNN Error (%)

0.999
0.10
1.010
1.00
0.997
0.30
1.000
0.00
1.000
0.00
1.000
0.00
1.000
0.00
1.000
0.00
1.000
0.00
1.000
0.00
1.000
0.00
1.000
0.00

EAC (s) CNN (s) Error (%)

1.000
1.000
0.00
1.000
0.731
26.90
1.000
0.720
28.00
0.337
0.352
4.45
0.343
0.370
7.87
0.324
0.287
11.42
0.298
0.312
4.70
0.324
0.351
8.33
0.306
0.265
13.40
0.333
0.346
3.90
0.340
0.342
0.59
0.307
0.356
15.96

Using data of error presented in table XI, we

can conclude minimum and maximum accuracy
value of CNN method when it used to assess
CCT value for case 2 (b) as follow:
- Minimum accuracy : 72,00 %

0.950
0.850
0.750
CCT

No
1
2
3
4
5
6
7
8
9
10
11
12

0.650
0.550
0.450
0.350
0.250
0

10

15

20

25

Data keCNN

EAC

Figure 15. Comparison chart of CCT for case 2 (b)

CONCLUSION
Committee neural network (CNN) proposed in this
research is used to classify transient stability status and
critical clearing time (CCT) of the system for several three
phase faults occurred on the system. Calculations using
equal area criteria were first carried out to determine
transient stability status of the system and CCT at several
generation patterns, fault locations, and transmission lines
to cut off. Furthermore, generation patterns, faulted buses,
and transmission lines to cut off are used as an input for
CNN. Each neural network expert which is constructing
CNN trained individually using backpropagation algorithm.
This proposed method tested using several generation
patterns for different fault locations and transmission lines
to cut off at Java Bali 500 kV multi machine electrical
power system. Results show that this method can achieve
better assessment when used to assess transient stability
status than CCT. The CCT values obtained by the CNN
were not consistently accurate, it is not like the transient
stability status which has good assessment accuracy. Results
from case 1 and case 2 show that when CNN tested using
untrained pattern to assess transient stability status, its
minimum accuracy are more than 90%, but when CNN is
used to assess CCT, its minimum accuracy are less than
80%. To improve the transient stability assessments
accuracy, it is recommended to add more learning patterns,
increase variation of the learning patterns, or add more
inputs for neural network.
REFERENCES
[1]. Imam Robandi, Desain Sistem Tenaga Modern, Andi, Yogyakarta,
2006.
[2]. Djiteng Marsudi, Pembangkitan Energi Listrik, Erlangga, Jakarta,
2005.
[3]. Reza Ebrahimpour and E. K. Abharian, An Improved Method in
Transient Stability Assessment of a Power System Using Committee
Neural Networks, International Journal of Computer Science and
Network Security, Vol. 9, No. 1, 2009.
[4]. Simon Haykin, Neural Networks a Comprehensive Foundation,
Pearson Prentice Hall, India, 2005.
[5]. Mauridhi Hery P. and Agus Kurniawan, Supervised Neural Networks
dan Aplikasinya, Graha Ilmu, Yogyakarta, 2006.

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

56

Proceedings of CITEE, August 4, 2009

[6]. I. M. El-Amin and A. A. M. Al-Shams, Transient Stability

Assessment Using Artificial Neural Networks, Electric Power
System Research 40, pp: 7-16, 1997.
[7]. N. I. A. Wahab and Azah Mohamed, Transient Stability Assessment
of a Power System Using Probabilistic Neural Network, American
Journal of Applied Sciences 5, (9): 1225-232, ISSN 1546-9239, 2008.
[8]. Hadi Saadat, Power System Analysis, McGraw Hill, Singapura, 2004.
[9]. Prabha Kundur, Power System Stability and Control, Mc. Graw Hill,
Singapura, 1994.

ISSN: 2085-6350

[10]. Boy Sandra, Design of Multi Machine Transient Stability

Assessment Application Software Using Equal Area Criteria
Method, Final Project on Electrical Engineering Department, Faculty
of Industrial Technology Sepuluh Nopember Institute of
Technology, Surabaya, 2009.
[11]. Saludin Muis, Teknik Jaringan Syaraf Tiruan, Graha Ilmu,
Yogyakarta, 2006.
[12]. J. J. Siang, Jaringan Syaraf Tiruan dan Pemrogramannya
Menggunakan Matlab, Andi, Yogyakarta, 2006.

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

57

Overcurrent Protection Coordination Due to Liquid Starter Effect on Large

Induction Motor
Dimas Anton Asfani, Nalendra Permana
Jurusan Teknik Elektro - FTI, Institut Teknologi Sepuluh Nopember
Kampus ITS, Keputih - Sukolilo Surabaya 60111
[email protected]

AbstractDuring motor startup, motor will draw

600-800% full load current. This condition will cause power
quality disturbance. There are several motor starting methods
which use to reduce motor starting current. Rotor resistance
method is one of starting method which common use to startup
the slip-ring induction.
Rotor resistance starting method reduces motor
starting current to 250% FLA. This method also increase
motor locked-rotor torque because rotor resistance addition
will change the slip of maximum-torque. There are two kind of
resistance, metal resistance and liquid resistance.
In this research, rotor resistance characteristic will
be applied then will be analyzed and examined as
considerations to determine the motor protective relays
setting . Results of this research indicate that the use of rotor
resistance affect the relays setting.
KeywordRotor Resistance Starting Method, Motor
Protective Relay.

I. Introduction
Induction motor is an electric motor which
commonly use for industrial application. Its advantages is
cheap, easy to control its speed and torque, and simple to
start [1]. During motor starting programs, it will draw
several times its full load current which will cause bus
voltage drop, that could cause motor speed variation and
nuisance tripping of protective relays [2].
Motor uses starter to reduce its starting current to
prevent the problem caused by starting program, then it will
not harm motors mechanical and electrical damage point..
The usage of motor starting method will change
motors current and accelerating time respond then we need
to do transient analysis every type of motor starter
Motor starting method is classified to four classes,
direct on-line starter, reduced voltage starting method ,
including wye-delta starter, autotransformer, and stator
resistance; rotor resistance for slipring motor, and power
electronic starting method, including VFD and soft starter
[3]
Reduced voltage starting method will reduce
motors terminal voltage during startup program then apply
full voltage when motor almost reach its nominal speed.
This method effectively reduce starting current but it also
reduce starting torque to square reduced voltage [4].
Wound rotor induction motor has stator wound like
the squirrel-cage induction motor, but its rotor wound
tapped out using slipring and carbon brush. It purpose to

add an additional resistance series linked with rotor wound

during motor startup program. After startup and motor
nearly reach its nominal speed, the rotor resistance will be
short-circuited by contact. Additional rotor resistance will
reduce starting current and increase motor starting torque.
There are two kind of resistance which commonly
use, metal resistance and liquid resistance. Metal resistance
use metal contact to control the resistance value. Further the
contact from input terminal, higher the resistance value. the
disadvantage is the metal contact at the risk of melt down
due to high starting current. The second type of resistance is
liquid resistance. It has principle that electrolyte immerse
two conductor bars. Higher level the electrolyte immersion,
lower the resistance value. liquid resistance has negative
thermal coefficient. If the temperature of electrolyte is
rising, the resistance value will decrease. It resticts the
frequency starting and minimum time between the starts [5].
The advantage is the resistance value of liquid resistance
decreased gradually and smoothly during the starts because
it depend on level of electrolyte immersion and no risk of
metal contact melt down

II. METHODOLOGY
The method that use in this research is modelling
liquid starter rotor resistance based on its resistance value
control characteristic and calculation for resetting
overcurrent relay.
A. Modelling liquid starter rotor resistance
First, modelling the response of torque-to-slip,
current-to-slip, and power factor-to-slip (Pf-to-slip). The
data is acquired from Tonasa IV cement industries, crusher
motor 402CR01M1, 1800 kW, 6.3 kV. Here is the acquired
data :
1. Stators resistance and reactance.
2. Rotors resistance and reactance.
3. value of external resistance measured data.
4. Synchronous motor speed
5. Nominal power output, nominal voltage, and
nominal power factor
there are several calculation to get the response of
torque-to-slip, current-to-slip, and power factor-to-slip, here
is the flowchart of calculation step :

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

58

Proceedings of CITEE, August 4, 2009

Step 1
Calculation thevenin
equivalent circuit parameter

( )=

i A

and,

Step 2
Calculation response torque-to-slip
motor starts

First step is calculation thevenin equivalent circuit

parameter to acquire thevenin voltage and thevenin
impedance. These parameter is calculated using equation (1)
and (2) [6].

Fig. 2.a. motor equivalent circuit

Fig. 2.b. thevenin equivalent circuit

v_th = v_phase * ( xm / (r12 + (x1 + xm)2) )

(1)

z_th = ((j*xm) * (r1 + j*x1)) / (r1 + j*(x1 + xm))

(2)

r1 and x1 is stator resistance and reactance, xm is

magnetism element reactance, and v_th dan z_th is thevenin
equivalent circuit voltage and impedance.
Second step to acquire response torque-to-slip
motor starts using torque equation [6] :

R th +

r2
s

( X th + x 2 )

((

3 * Z in

(4)

(3)

(6)

The calculation of the torque-to-slip, current-toslip, and Pf-to-slip consider the characteristic of liquid
starter rotor resistance that the resistance value will decrease
gradually and smoothly. Tonasa IV crusher motor
402CR01M1 uses 12 step of decreasingly resistance value
then there will be 12 response for every rotor total
resistance value, rotor resistance add up with external
resistance, then will be compounded for every step of motor
speed from start condition (slip = 1) until synchronous
speed (slip = 0).
The response of torque-to-slip, current-to-slip, and
Pf-to-slip for motor starting condition will be used as motor
model characteristic then we do motor starting analysis
using ETAP. We will get inrush current during motor
starts. The inrush current will be analyzed and examined as
consideration to determine the setting of motor overcurrent
relays.
B. Motor Overcurrent Relays Resetting
We need to calculate the setting to get overcurrent
relays setting that could protect motor from overload and
short circuit fault using equation :
1. Pickup current tap
(7)
1.05 * I FL < I pp < 0.8 * I SC min
Tap( sec.) = I PP / nCT

( )

Tap I n = Tap ( sec .) / I n sec .CT

2.

(8)
(9)

Time dial inverse

IEEE Very Inverse [8]

td ( I ) =

T
3.922

2
0
.
138
I
1
IS

()

(10)

While,
T
= 0.1
I
= ISC-Max
Is
= arus pickup

q1 is number of phase, s is synchronous speed in second

(rps), Rth and Xth is thevenin equivalent circuit resistance
and reactance, r2 and x2 is rotor resistance and reactance.
Third step to acquire response current-to-slip
motor starts using current equation [7] :

ISSN: 2085-6350

Im I
Pf = Cos arcTg

I
Re

Fig. 1 Flowchart of modelling liquid starter rotor resistance

r2
+ jx 2
jx m
(5)
s
i(A) is starting current in Ampere and Zin is input
impedance in Ohm ().
Pf-slip motor starts response could be acquired
using Pf equation [7] :

Step 4
Calculation response Pf-to-slip motor
starts

T =

rating

Z in = r1+ jx1 +

Step 3
Calculation response current-to-slip
motor starts

2 r
q 1 V th 2

3.

Definite tap setting

I Def ( sec .) = I SC min / nCT

(11)

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

59

( )

Tap definite I n < I Def ( sec .) / I n sec . CT

20.00
15.00
10.00
5.00
3.33
1.66
0.00

(12)

III. CASE STUDY

Tonasa IV cement industry sub-station 1 sub-group
2 has crusher motor 402CR01M1, which use as liquid
starter motor model, has data that shown in table 1.

372.1
456.5
379.5
361.8
274.7
152.6
19.4

52.6
50.1
56.9
67.0
76.8
86.7
47.2

Torque response is calculated using equation (3)

for every level of external resistance value. The current
response is calculated using equation (4) and Pf using
equation (6).
From Table 1, we will get torque-to-slip, currentto-slip, and Pf-to-slip curves that shown in figures below :

Table 1
Motor data 402CR01M1
Nominal voltage
6300 kV
Synchronous speed
750 rpm
Nominal Pf
0.8
Stator resistance
0.1534
Stator reactance
0.410
Rotor resistance
0.0270
Rotor reactance
0.410
Stator winding/rotor winding
(approximate)
1
Measured value of external resistance
No.

No.

No

No
1
1.7
4
0.81
7
0.38
10
2
1.09
5
0.65
8
0.27
11
3
0.93
6
0.52
9
0.15
12

148.2
118.4
146.3
150.9
161.5
133.2
11.3

L iquid S tarter T orque

% F LT
180.0

L iquid s tarter Torque

1.7
1.09
0.93
0.81
0.65
0.52
0.38
0.27
0.15
0.077
0.023
0

160.0
140.0
120.0

0.077
0.023
0

100.0
80.0
60.0
40.0
20.0
0.001

6.668

13.334

20.001

26.667

33.334

40.001

46.667

53.334

60.000

66.667

73.334

80.000

93.333

86.667

0.0
100.000

The calculation using equation (1) until equation

(6) the we will get the motor response using liquid starter
that shown in table 2.

% S lip

Fig. 3. Motor starting torque response using liquid starter

1.09
0.93

400.0

0.81
300.0

0.65
0.52

200.0

0.38
100.0

0.27
0.15

% slip

0.100

5.000

10.000

15.000

20.000

25.000

30.000

35.000

40.000

45.000

50.000

55.000

60.000

65.000

70.000

0.0
75.000

52.4
51.5
53.2

1.7
500.0

80.000

333.0
417.5
403.9

L iquid s tarter C urrent

600.0

85.000

% Pf
87.3
85.8
83.2
81.9
81.5
81.1
80.9
80.6
80.2
79.5
78.2
77.6
76.2
75.6
73.8
72.6
70.4
69.0
66.5
65.3
61.8
59.2
57.6
58.6
57.3
55.1
54.0
55.3
54.5

90.000

% I FL
145.8
143.6
139.1
136.9
196.0
192.8
186.4
209.5
202.3
198.7
214.7
210.6
202.4
198.2
227.4
222.6
212.7
247.7
236.2
230.3
276.5
269.5
254.9
247.4
295.9
286.8
267.8
368.0
345.4

700.0

95.000

Slip
%Torque FL
100.00
129.8
98.33
128.6
95.00
126.1
93.33
124.8
90.00
150.6
88.33
149.6
85.00
147.6
83.33
154.2
80.00
152.4
78.33
151.4
75.00
155.4
73.33
154.5
70.00
152.4
68.33
151.2
65.00
157.8
63.33
156.9
60.00
154.9
58.33
160.4
55.00
159.1
53.33
158.3
50.00
161.5
48.33
161.5
45.00
160.9
43.33
160.3
40.00
160.8
38.33
161.3
35.00
161.5
33.33
149.3
30.00
154.3
Continuation of Table 2
28.33
156.5
25.00
134.3
23.33
138.9

% IFL

100.000

Table 2
Motor response using liquid starter plot

0.077
0.023
0

Fig. 4. Motor starting current response using liquid starter

Pf

Fig. 5. Motor starting Pf response using liquid starter

Then, data from Table 2 will be analyzed using motor

starting analysis thus we obtained the motor starting current

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

60

Proceedings of CITEE, August 4, 2009

using liquid starter. Fig. 6 will show the result of

simulation.

Relay code
ISC-Max (A)
ISC-Min (A)
IFL Motor (A)
Tap Setting
Inverse
Time Dial
Tap I >>
Delay

440SG2F2
resetting
22186
10326
206.41
0.63
0.75
IEEE Very Inverse
12
0.4
5
5
0.1
0.1

Existing

Fig. 6 Motor 402CR01M1 starting current

Data of time duration-to-current can be determined

by conducting a sampling rate for each current level and
then determined the length of time for the current level. Fig.
7 will illustrate the process of determination time durationto-current data
% I FL

125 % I FL
5.94

Fig. 7 Illustration of time duration-to-current

For example, current level 125% full load current

(FLA), then the result shown in Table 4.3 that starting
current above 125% FLA or 125%

1800
=258.1
3 6.30.8

A starts from t = 1 sec until t = 5.94 sec then duration of

125% FLA is 5.94 - 1= 4.94 sec.
Motor starting current using liquid starter has
lower current magnitude and shorter acceleration time
because external resistance addition will increase the
starting torque characteristic. due to significant changing on
starting current and duration, the setting of protection relay
will need to be recalculated. Table 3 shows the existing
condition and its resetting recommendation
Motor starting current will be analyzed for its
overcurrent relays setting. Fig. 8 shows motor
402CR01M1s starting current and existing relays setting.
The setting includes tap setting calculation using equation
(7), (8), and (9), time dial inverse calculation using equation
(10), and tap definite setting calculation using equation (11)
and (12)

Fig. 8 Direct on-line motor starting current and existing relay setting

The examination result of fig. 8 is the existing

relay setting optimal protect the motor only if motor use
direct on-line starting. When the motor uses rotor resistance
starter, the motor starting current is reduced and motor
accelerates faster then the existing relay setting no longer
optimal protect the motor. Fig. 9 shows the motor starting
using liquid starter and the recommended setting that
provides optimal protection to motor during starting and
protect the motor from overload and short-circuit fault.

Table 3
Existing and recommendation motor overcurrent relay setting

ISSN: 2085-6350

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

61

[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]

[18]
Fig. 9 Liquid starter rotor resistance motor starting current and
recommended relay setting

IV. CONCLUSION.
The usage of motor starter purposes to reduce
starting current then will reduce power quality problem.
Motor starter will change the torque, current, and power
factor motor characteristic during induction motor starts. It
will affect to relays setting then the setting need to be
recalculated for specific motor starter. The existing setting
will be replaced by the new setting, that consider the usage
of motor starter, then the relay remains closed during motor
starting program and also protect the motor from overload
and short-circuit fault.

[1]
[2]
[3]

[4]
[5]
[6]
[7]
[8]

[19]

[20]
[21]

-,Sepam Series 20 Merlin Gerin,Schneider

Electric, 2003
Stephen J. Chapman, Instructors Manual to
Accompany Electric Machinery Fundamentals,
McGraw-Hill, 2004.
Abdullah Saeed Al-Amoudi, Air conditioners
Peformance Using Soft Starter, King Fahd
University of Petroleum & Minerals, 2003
-, QFE Manual, Fairford Electronic, 2008. Hal. 63
Irwin Lazar, Electrical System Analysis and Design
for Industrial Plants, McGraw-Hill, 1980.
R Wahyudi, Diktat Kuliah Sistem Pengaman
Tenaga Listrik, 2008
-, Rotor Starters for Slipring Motors, Pape &
Olbertz, 2008.
-, Slip Ring Induction Motor Starter. ABB
Australia Ply Limited, 2008.
-, IEEE Recommended Practice for Protection and
Coordination of Industrial and Commercial Power
Systems. Industrial and Commercial Power Systems
Committee of the IEEE Industry Applications
Society, 1986.
David C. Yu, Protective Device Coordination
Enhancement for Motor Starting Programs. IEEE
Transactions on Power Delivery, vol. 20, no. 1,
January 2005.
Pedro Jiminez dkk, Motor Starting Study for Large
Motors. Case: VALCOR PDVSA Project. IEEE
PES Transmission and Distribution Conference and
Exposition Latin America, Venezuela, 2006
Viv Cohen, Induction Motors - Protection And
Starting. Circuit Breaker Industries, South Africa,
2000
Nathan Schachter, Experience with synchronous
and slip ring induction motors driving cement mills.
IEEE Cement Industry technical Conference, 1998

V. REFERENCE.
Soebagio, Teori Umum Mesin Listrik,Srikandi,
2008.
Soebagio, Diktat Kuliah Mesin Arus Bolak-Balik,,
2008.
David C. Yu dkk, Protective Device Coordination
Enhancement for Motor Starting Programs, IEEE
Transactions on Power Delivery, vol. 20, no. 1,
January 2005.
Andy Smith, Electro-Mechanical Method of
Starting Induction Motors, Somar International Ltd.
Technical Document, 2004.
John Larabee dkk, Induction Motor Starting
Methods and Issues, Siemens Energy &
Automation Inc., 2005.
-, Resistance Motor Starter, ABB Australia Ply
Limited, 2008.
A. E. Fitzgerald dkk, Electric Machinery Third
Edition, McGraw-Hill Book Company. 1971.
K A Walshe, AC Motors Rev:A.

Conference on Information Technology and Electrical Engineering (CITEE)

VI. BIOGRAPHIES
Dimas Anton Asfani was born in
Ponorogo, Indonesia, on 5
September 1981. Graduated from
Master of Engineering Sepuluh
Nopember Institute of Technology,
Surabaya Indonesia, with Cum
Laude. He is also lecturer of
Electrical Engineering Department in
same institute.
Nalendra Permana was born in
Surabaya, Indonesia, on 20 April
1988. He is student of ITS Electrical
Engineering Department, Sepuluh
Nopember Institute of Technology,
Surabaya Indonesia.

ISSN: 2085-6350

62

Proceedings of CITEE, August 4, 2009

Analysis of 20 kV PLN Relay Protection Typical

Setting in Industrial Customer
Dimas Anton Asfani, Iman Kurniawan
Department of electrical engineering, Sepuluh Nopember Institute of Technology
Surabaya 60111, Indonesia
e-mail : [email protected]

Adi Soeprijanto
Department of electrical engineering, Sepuluh Nopember Institute of Technology
Surabaya 60111, Indonesia
e-mail : [email protected]
Abstract The need of industrial customers for electrical
energy has increased and demanded a continuous service from
PLN. The continuity of electrical service can be optimized by
coordinating proper protections for both PLN and its
industrial customers. The PLN regulation states that setting
relay to limit overload for industries is adjusted to the
electrical capacity of customer contract; whereas PLNs over
current relay protection is coordinated with customers over
current relay protection based on the curve of IEC Standard
Inverse.
In some industrial cases, setting over current relay cannot
be coordinated well because PLNs protection for over load
relay and over current relay is coordinated with inverse and
instant curves. Consequently if a short circuit occurs in a
customers electrical installation, PLNs relay trips. This study
is intended to recommend PLN to use of two inverse curves for
setting over load relay and over current relay. This setting is
coordinated with the customers relay with regard to
customers short circuit current and the electrical capacity of
customer contract.
Keywords continuity, over current relay, over load relay

INTRODUCTION
PLN as electricity suppliers in the country to a restriction in
the power of their customer. For customers 20 kV medium
voltage limiting usage is charged with the circuit breaker
equipped by over load relay three phase with a time
characteristic that the formula refers to the cold start
characteristics thermis from over load relay and adjusted
according to the current nominal power connected. PLN
limit on the condition of four times nominal current settings
over load relay is coordinated with the over current relay
[1].
Customers in this category are industries that require a
reliable protection in their electricity system with a purpose
when there is short circuit fault no interference with the
production path. Relays settings are expected to qualify
sensitive, selective, reliable, and react quickly. One method
is to obtain reliable system is to enable relay as the main
and backup protection. The backup will work when the
main function of the failure occurs when the fault
happen[2]. To fulfill these functions relay as main

ISSN: 2085-6350

protection must set earlier than the backup relay. With this
setting at the time fault occurred, the main protection relay
will trip at the first time to eliminate the impact of the
widespread disruption as an effect of fault. Separate the
relays functions become the main and backup protection
into consideration in applying the concept of areas
protection system in which the absolute reliability is
required [3]. PLN apply relays function as the limiter and
protection. But according to the concept of relays function
as the primary and the backup, so PLNs relay is the lastt
back up which is work if customers relay not function
when fault happen on the electrical system customers.
MAIN THEORY
One of over current fault of electric power system is a
three phase short circuit fault. Knowing the nominal of
current when short cicuit fault happen can be used as
reference to coordinate relay protection settings so there
surrent not damage the equipment and system stability.
Practical, to calculate the flow of short circuit current in
the middle voltage distribution system can be made as
follows:
a. Three phase short circuit falut [3]
Isc 3 ph =

VL - N
Z1

Isc 3 ph =

MVA f

Isc 3 ph =

VL - L

MVAsc =

(1)
(2)

3 kV

(3)

3 Z1
MVA b

(4)

Z1

b. Phase to phase short circuit fault[3]

Isc 2 ph = 0.866 xIsc 3 ph

(5)

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

63

The methode to analyze can see on the next flow chart

as follows :

DELAYS TIME SETTING

Over current relays setting has vital function on the
relays protection coordination. Relays setting can be made
by time-based setting, current setting or a combination of
both [8].
At the time of setting we know about the delay time
(t). The difference between the minimum working time
main relay and backup relay is 0.3 - 0.4 sec [3].
tset = t + t

(6)

where: t is the time difference of 0.3 - 0.4 sec, and t is

the time setting main relay on the feeder, and tset is time
setting back up relay on the feeder.

Figure 1. Setting relays coordination with the delay time

An example in the picture above is when the t (time) in

the relay C bus = 0.1 second, the time delay (t) 0.3 second
left, then obtained the time on the relay in the B bus = 0.1 +
0.3 = 0.4 second, the time in exit feeder (A bus relays) = 0.4
+ 0.3 = 0.7 second, and the time in the incoming feeder =
0.7 + 0.3 = 1 second.
While for the currents setting and the combination of
time and current, we know about of Inverse Definite
Minimum Time (IDMT), These relays characteristics has
some parts of the inverse and definite [4]. Relays setting
also must consider the error pick-up in accordance with
British Standard Pick Up = 1.05 - 1.3 nominal currents [4].

B. Thermal overload relay

It have a function to protect the load from the electric
motor from over burden. Some cause more of a burden,
such as starting current too large, the mechanical motor load
is too large, the motor is stopped suddenly, one of the
opening phase of the motor three-phase line, the occurrence
of short circuit fault, etc.
Time characteristic of over load relay reference to cold
start formulas from thermis characteristic over load relay as
follows

(I )2 menit
(I )2 (kxIs )2

x ln

Figure 2. Analyze metodology flowchart

A. Research Location
Locations objects are industrial customer 20 kV middle
voltage which located on North Surabaya APJ PLN and
South Surabaya APJ PLN.
TABLE I RESEARCH LOCATION
Customers

Capacity
(kVA)

Pasar Atum

Location

PLN Area

3465

Jl. Bunguran No. 45 Sby

Sby. Utara

T. Plasa

13840

Jl. Basuki Rahmat No. 812 Sby

Sby. Utara

Juanda

9500

Jl. Ir. H. Juanda No.1 Sdrjo

Sby.
Selatan

Pelindo III

13840

Jl. T. Perak Timur No. 620

Sby

Sby. Utara

B. PLN Impedance System

Impedance values in the network PLN which is
influence of short circuit current came from impedance on
substation, transformer, and from cable which as the
channel between the source and the bus fault.

.
(7)
where t is time in minutes, is a constant thermis, ln is the
natural logarithm, I is the load current, k is a constant
(1.05), and I is the currents setting of relay [8].

METODOLOGY AND CASE STUDY

Figure 3. Impedance on the radial system

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

64

Proceedings of CITEE, August 4, 2009

TABLE II. IMPEDANCE SYSTEM SUBSTATION [6]

Substati
on

Voltage
(kV)

Pasar Atum

Undaaan

T. Plasa

Customers

Impedance
R1 (pu)

X1 (pu)

150

0.0014381

0.0120965

Sawahan

150

0.001155

0.0111775

Juanda

Waru

150

0.001496

0.0126186

Pelindo III

Perak

150

0.0030887

0.0178338

The table above is impedance datas substation

supplying customers who become objects of research which
is shown by the large of resistance and reaktance in per unit
(pu).
While for the impedance of transformer, based on
SPLN 8-5: 1991, the values as follows :
TABLE III IMPEDANCE OF TRANSFORMER
SEPARATE WINDING ON THE SUBSTATION

Substation
Undaaan
Sawahan
Waru
Perak

Transformers
Capacity
(MVA)
60
50
60
50

WITH

C. Setting divider Relay PLN 20 kV.

One of the limitations of power conducted by PLN for the
connection in the middle voltage network is used a medium
voltage circuit breaker equipped with relay with over load
current characteristic refering to the formula of cold start
from thermis overload relay characteristics [1]. Limit and
the measurement of medium voltage 20 kV applied in PLN
is as follows:
TABLE VI. LIMITS AND MEASUREMENT 20 KV PLN MIDDLE
VOLTAGE CUSTOMER [1]
Voltage System
(kV)

Capacity
(kVA)

Current Divider (Ampere)

20

2.180
2.770
3.465
4.330
5.540
6.930
8.660

63
80
100
125
160
200
250

TWO

Impedance(%)
12.7
13.3
12.3
12.5

In accordance with the formula of cold start from thermis

over load relay characteristics, it have tripping characteristic
for assessment as follows:
TABLE VII. PREFERENCES SETTINGS RELAYS DIVIDER 20 KV
PLN CUSTOMERS [1]
Current

Time to trip

For the customers transformer with biggest capacity, the

impedance are

1,05 x In

Trip before 60 minutes

1,20 x In

Trip before 20 minutes

TABLE IV. IMPEDANCE OF TRANSFORMERS CUSTOMERS

1,50 x In

Trip before 10 minutes

Coordinated with short circuit current
protection (OCR)

4,00 x In

Customers

Transformers
Capacity
(MVA)

Impedance(%)

Pasar Atom

3000

T. Plasa

5000

Juanda

4500

Pelindo III

2000

6.25

In is a nominal current of customers according to the PLN

contract. The maximum currents uses of customers is
limited to four times nominal currents.
TABLE VIII. EXISTING SETTING CONDITION 20 KV PLN RELAY
DIVIDER

For the conductor as a reference standard is SPLN PLN 64:

1985, impedance values as follows

Customer

kVA

Curv

Tap

Tms

Inst.

Del.

CT

Pasar
Atom

3465

IEC
EIT

0.21

100/5

T. Plasa

13840

LTI
Alst

0.5

0.15

400/5

Juanda

9500

UIT

0.7

0.8

2.7

0.22

400/5

Pelindo III

13840

IEC
EIT

1.1

0.13

400/5

TABLE V IMPEDANCE CONDUCTOR AAAC [7]

Conductor Measure
Width Area (mm2)
Core
16
7
25
7
35
7
50
7
70
7
95
19
120
19
150
19
185
19
240
19

Impedance
( / km )
2.3569 + j 1.5480
1.6760 + j 1.1804
1.3192 + j 1.3262
1.0089 + j 1.2364
0.8245 + j 1.2258
0.7033 + j 1.1782
0.5647 + j 1.0697
0.5347 + j 1.0597
0.4611 + j 1.0129
0.3821 + j 0.9702

Capacity

CALCULATION AND ANALYZE

ISSN: 2085-6350

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

65

A.
Analyze of impedance value radial electrical system
from PLN to customers
The first step in calculation of short circuit current is
calculate the value of impedance on the electrical system is
which is trip. Here is a single line diagram Pasar Atum, and
the arrow shows the calculation current short circuit
analysis conducted

Substation

0.012

Substation
Transformer

0.2125

Conductor

0.623

Z total

Cable

0.02

Customer
Transformer

0.23

3.1675

Figure 5. Reaktance diagram Mall Pasar Atum when short circuit current
fault happen on the low voltage bus transformer with a biggest power
capacity

From the above reaktance diagram total impedance value

obtained is 3.1675 pu. With the normalization KVbase to
0.4 kV, the actual value obtained was 0.00508 .

B. Analyse of Short Circuit and Relays Coordination on

Low Voltage Transformer Customers.

Z1 = 0.00508

I f 3 ph =
Figure 4. Single Line Diagram Mall Pasar Atum

Analysis is done on the low voltage transformer with a

largest capacity on the customers, because low voltage bus
transformer will result in the largest currents at the time of
short circuit fault happen compared to the other transformer
that has smaller power capacity.
Power supplier on the electrical system Mall Pasar
Atum is Undaan substation. These substation has resistance
and reaktance as described above. Transformer substation
has 60 MVA capacitys with impedance about 12.7%.
Conductor between main substation bus to interference
substation and from it to incoming customers bus will also
have impedance in accordance with the length. Transformer
on the customer side also has the impedance of course also
affect.
Results of MVAbase normalization with 100 MVA
values obtained Impedance (in pu) of the electricity system
Pasar Atum as follows:
TABLE IX. IMPEDANCE OF PASAR ATUM ELECTRICITY SYSTEM
Equipment
Substation
Substation Transformer
Conductor
Cable
Customers Transformer

Impedance (pu)
0.0120
0.2125
0.6230
0.0200
2.3000

Reactance diagram from Pasar Atum electricity system

when short circuit fault happen can see as follows

If 2

ph

L L

3 * Z1
0.4
3 * 0.00508

= 45.48 kA
= 0.866 * I sc 3 phasa
= 0.866 * 45.48
= 39.39 kA

Using the same way short circuit analysis is done on another

location and obtained results as follows:
TABLE X. SHORT CIRCUIT CURRENT AT ON LOW VOLTAGE BUS
CUSTOMERS TRANSFORMER
Customer

Bus ID

Volt
(kV)

Iscmax
(kA)
4 cycle

Iscmax
Iscmin
(kA)
(kA)
30 cycle

P. Atum

PA - 6

0.4

46.97

45.48

39.39

T. Plasa

MVDP1

6.6

10.84

10.36

8.97

Juanda

MPH-M11

6.6

4.87

4.73

4.09

Pelindo
III

Out - TPS4

0.4

37.75

37.27

32.28

From the results of the calculation short circuit current,

we can compare with the nominal current in each location,
so that can be counted tides that need to be secured in order
not to damage equipment and disrupt the stability system.

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

66

Proceedings of CITEE, August 4, 2009

TABLE XI. COMPARISON BETWEEN NOMINAL CURRENT AND

SHORT CIRCUIT CURRENT ON MEDIUM VOLTAGE 20 KV BUS
INDUSTRIAL CUSTOMERS
PLN Contract
Customer

Capacity

SC on LV

In

Volt

Isc Max

Konv 20
kV

(kVA)

Amp

kA

(kV)

(kA)

(kA)

Pasar
Atom

3465

100

0.1

0.4

46.97

0.93

T. Plasa

13840

400

0.4

6.6

10.84

3.59

Juanda

9500

275

0.27

6.6

4.87

1.61

Pelindo III

13840

400

0.4

0.4

37.75

0.75

Need to be
compliteed

Mall Pasar Atum using relay as the protection equipment

on their electrical system, there are on the side of the
incoming and outgoing PLNs source and outgoing to the
each transformer.
TABLE XII. EXISTING SETTINGS CONDITION PASAR ATUM
INTERNAL RELAYS
Relay
MV-01
Internal
MV-02
Internal
MV-03
Internal
MV-04
Internal

CT
Ratio
250/5
100/5
75/5
75 /5

C Resetting 20 kV PLNs Relay in the Customer

Curve
IEC
Inv
IEC
Inv
IEC
Inv
IEC
Inv

Tap

Tms

Inst.

Del

0.66

0.10

1.3

0.14

0.70

0.11

1.4

0.17

0.63

0.11

1.3

0.14

0.63

0.10

1.3

0.26

By knowing short circuit current on the customer side,

then we will be able to analyze the coordination of the
protection settings on the customer and PLN. Existing
condition PLNs relay setting and customers relay as shown
in the coordination curve as follows not ideal. Especially
on the grading / lag working time between relay. With t
(delay) less than 0.3 seconds, it can be ascertained that
when short circuit fault happen on the customer, then PLNs
relay also join to trip, so customer electrical system will
totally trip. Not even close the possibility of these condition
make main substation relays also trip and black out will be
more knowledgeable.
In addition, due to the condition of short circuit current
is greater than four times nominal current (In), it can be
ascertained at the time it happens customer electrical side
will totaly trip.

ISSN: 2085-6350

Figure 5. Existing coordination curve rele PLN and Pasar Atum relays low
voltage bus PA-6

Coordination of PLNs 20 kV use the type of IDMT

Curves (Inverse Definite Minimum Time), it consist of two
curve setting on a relay. IDMT type is also known by code
51/50. 51 is the code to ANSI standard over current relay
with inverse setting, while 50 is the code to ANSI standard
over current relay with instant settings [3].
On the fact it often make it customes difficult to adjust
their internal settings with PLNs relay. Although as a
power divider, but on the Tarif Dasar Listrik (TDL) 2003
also shown that the condition four times current nominal
(In) is coordinated with the over current relay, so it have
two function as power divider and also protection relay.
PLNs relay as protection that need to be coordinated with
both the customers internal relays setting, so that when the
short circuit fault happen in the customer system, then
PLNs relay will working as a back up relay with a certain
time
delay
To analyze 20 kV PLNs relay setting at a coordination
protection system must be in the ideal condition especially
for the internal customers relay. PLNs relay setting
suggested to used two inverse curve (51/51), which consists
of a over load relay (OLR) curve and over current relay
(OCR) curve. Over load relay curve setting is still in
compliance with provisions of the setting standard on TDL
2003, there are on the condition 1.05 times nominal current,
1.2 times nominal current and 1.5 times nominal current.
While over current relay curve setting is still on to the
British Standard Pick Up = 1.05 - 1.3 times nomnal current
[4]. But the starting point is 1.5 times the nominal current.
Range is taken between 1.05 1.3 is 1.2. So for the over
current relay is set at 1.5 x 1.2 = 1.8 times nominal current.
Selected curve is IEC Standard Inverse [4].

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

0,14

X (T
0 , 02
If

1
I Set

67

T 0.4849
Selected = 0.9

)
(8)

While t is the time setting to trip, If is maximum short

circuit current as a setting reference relay, Iset is relays
current setting, and T indicates time delay operation (Tms)
of the relay.

Figure 7. Coordination curve resetting rele 20 kV PLN relay using two

inverse curve (51/51) with internal Pasar Atum relay

The same method to analyze protection coordination setting

20 kV PLNs relay on the other research locations. By
setting 20 kV PLNs relay refers to the short circuit current
in the low voltage largest capacity, then it can has good
coordinated.with customers relays.
TABLE XIII. RESETTING RESULT FOR 20 KV PLNS RELAY AT
THE MEDIUM VOLTAGE INDUSTRIAL CUSTOMERS
Figure 6. Existing setting coordination curve with two inverse curve
(51/51) for 100 Amp nominal current customers which is reference to four
times current nominal (In)

To set 20 kV PLNs relay Pasar Atum as follows :

Power Capacity = 3465 kVA
Nominal Current (In) = 100 Ampere
= 4 x 100 = 400 A
Four times In
Relay Current Setting (Iset)
= 1.8 x In = 1.8 x 100 = 180 A
Iscmax 4 cycle Bus PA-6 = 939.3 Ampere
nCT = 100/5
Tap = 3465 kVA
Type = MG SEPAM 1000+
Curve = IEC Standart Inverse [8]

0.7

=
T = 0.4849

T
0.14

X
0 . 02
2
. 97
939.3

180

0.14
0.09977

PLN Contract
Custom

Resetting 20 kV PLNs Relay

Cap

In

kVA

P.Atom

3465

100

SEPAM
1000+

T Plasa

13840

400

ALSTOM
P122

275

SEPAM
1000

Juanda
Pelindo
III

9500
13840

400

TAP (51) OLR

TAP (51) OCR

Curv

(x In)

Curv

(x In)

100/5

Cold
Start

1.05

IEC SIT

1.8

0.9

400/5

Cold
Start

1.05

IEC SIT

1.8

0.13

400/5

Cold
Start

1.05

IEC SIT

1.2

0.55

SIEMENS
400/5
7SJ600

Cold
Start

1.05

IEC SIT

1.8

0.035

Type

CT

Tms

CONCLUSION

From the calculation and analyze 20 kV PLNs relay

on the industrial customer and also analyze short circuit
current on the low voltage largest capacity in the customer,
then the conclusion are there are some miss coordination
setting between 20 kV PLNs relay with customer internal
relays setting. This is caused by short circuit current greater
than four times current nominal, while power divider 20 kV
PLNs relays setting own tend to trip them instantly on the
four times current nominal (In). This condition make
customer difficult to set their internal relay because the
reference setting are maximum short circuit current. But it
can be done with good conduct, by set 20 kV PLNs relay
with reference to maximal short circuit current which
happen at the low voltage largest capacity transformer bus
on the customers. In addition, coordination rele PLN 20 kV

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

68

Proceedings of CITEE, August 4, 2009

PLN relay coordination conducted with a time delay (t) in

accordance to grading standards in regulation time relays
setting, especially for electrical systems that have a typical
form..
REFERENCES
[1].
[2].
[3].

PT. PLN, Keputusan dan Edaran Tarif Dasar Listrik,

Januari 2003.
P. M. Anderson, Power System Protection, New York :
McGraw-Hill, 1999.
American
National
Standards
Institute,
IEEE
Recommended Practice for Protection and Coordination of
Industrial and Commercial Power System, IEEE Std 2421986

ISSN: 2085-6350

[4].
[5].
[6].
[7].
[8].
[9].
[10].
[11].
[12].
[13].

GEC Alsthom, Protective Relays Application Guide,

Stafford, England, 1987
Sunil. S. Rao, Switch Gear and Protection, Khanna
Publishes, 1980.
Arus Hubung Singkat 2007, www. bops.pln-jawa-bali.co.id
SPLN 64 : 1985, Petunjuk Pemilihan dan Penggunaan
Pelebur pada Sistem Distribusi Menengah
Merlin Gerin, Sepam 1000+ Substation Busbars
Alstom, Micom P120 / P121 / P 122 / P 123 Overcurrent
Protective Relays from V4 Software Version.
Merlin Gerin, Sepam 2000 Substation Busbars
Siemens, Numerical Time Overcurrent Protection and
Thermal Overload Relay.
SPLN 8-5 : 199, Kapasitas Hubung Singkat Transformator
SPLN 52-3 : 1983, Pola Pengaman Sistem Bagian Tiga,
Sistem Distribusi 6 kV dan 20 kV

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

69

The Electrical Energy Calculation Based Upon The Voltage Measurement

of a High Speed and Magnitude of Multiple Impulse Currents Produced
by an Impulse Generator Implemented to ZnO Block.
Haryono T.*), Sirait K.T.**), Tumiran*), Hamzah Berahim*),
*)

Dept. Of Electrical Eng., Faculty of Eng., Gadjah Mada University

Address: Jl. Grafika 2, Fakultas Teknik UGM, Yogyakarta, Indonesia
**)The School of Electrical and Informatics Engineering, Bandung

Abstract
An electrical equipment, such as a distribution
transformer, in its operation, should have an overvoltage or
a lightning protection. If a lightning strikes some electrical
distribution lines, a very high voltage produced endangering
the electrical equipment connected to the distribution lines.
In order to make the equipment in a safe condition, an
electrical arrester should be connected parallel to the
equipment in the distribution lines. The purpose of the
connected arrester is to limit the voltage due to the lightning
strike in a value that doesnot make any damage to the
protected electrical equipment. A kind of the electrical
arresters, ZnO arrester, consists of several arrester blocks of
zinc oxide material. In this experiment, it was shown the
ZnO voltage measurement method required enabling
impulse energy calculation to be done. An arrester will be
damaged if the energy received from the lightning strikes, in
this case multiple impulse current strikes, having high speed
and high magnitude, is more than the maximum of impulse
energy allowed. For 500 A impulse peak did not give bad
effect to the ZnO arrester block because the accumulated
impulse energy received by the ZnO block due to the 5
concecutive impulse currents was less then 50 J/cm3. The
limit is 200 J/cm3.
Key Words: Multiple Impulse, high speed and high
magnitude,multiple impulse current generator

Rated frequency
Nominal discharge current
Rated short circuit withstand current in kA.

To obtain the current flowing value through the arrester

and its voltage across it, Darveniza [2] used a digital storage
oscilloscope to record voltages and currents and then they
were manipulated to produce the energy absorbed by the
arrester. According Lat M.,V.[3], the calculation of energy
absorption is based on numerical integration of current and
voltage waveforms.

2. Multiple Impulse Current Generator

In order to simulate lightning current strikes that happen
several times in a certain duration of time, an impulse current
generator[5],[6],[7],[8],[9] consisting of 5 main capacitors was
constructed as in the picture shown below:
Diode

Brass rod
RC

Bras
plate

In any electrical system, it is important to provide all

electrical equipment in a safe operation condition. Therefore,
various protection systems are designed and implemented for
the electrical equipment, for example, over current
protection, under and over voltage protections, and other
kinds. Non linear metal oxide arrester[1], is one kind of some
electrical equipment protection against lightning strike. Any
arrester will cut a steep increase in voltage due to a lightning
strike in a safe voltage value, which is known as its residual
voltage. In other words, the arrester absorbs a greater part of
the lightning strikes energy and releases it to ground.
However, the capability of an arrester allowing the lightning
energy released to ground is limited to the arrester
specification as written on its specification as follows[1]
Continuous operating voltage
Rated voltage

Transformator
220/13200 Volt

Volta
-ge
devid
er 2

ZnO

compuscope

Current
shunt

1. Introduction

220
volt

R = 210 kohm

shield

Voltage
Devider 1
Grounding

Figure 1 Coaxial Construction Impulse current

Generator
Informations shown in Fig.1 are represented as shown
below:
C
= main capacitor, 8 F, 20 kV
RC
= Rogowski Coil, for
measuring current
ZnO
= one block of ZnO arrester
Compuscope
= computer oscilloscope for

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

70

Proceedings of CITEE, August 4, 2009

ZnO voltage and current

recording
voltage devider = for reducing the voltage that
is going to be measured
Current shunt
= for measuring the current
In order to minimize the inductance, the impulse current
generator, Fig.1, was developed on the coaxial construction
basis. According this way, the impulse current generator
main capacitor was located in the middle of construction, so
was the RC (Rogowski Coil), and the ZnO block. On top of
the main capacitor, it was located a pair of trigger electrodes
connected to top brass circular plate. By means of several
brass rods, the top brass circular plate was connected to the
bottom brass circular plate. A hole at the center of the
bottom circular plate is made for a current shunt device for
measuring the voltage across the current shunt. Then current
flowing through the ZnO block can be obtained by the
division of the voltage by the current shunt resistance, which
is 0,029 ohm. When the main capacitor, consisting of 45
single capacitors and their rating specification of 470 F
450 V in series connection or their measured total
capacitance of 8 F, was charged using DC voltage source
provided, and the sphere trigger was triggered, the impulse
current started to flow from the main capacitor going up
toward the top brass circular plate and going down through
all brass rods reaching the bottom brass circular plate.
Futher, the current flew through the current shunt back to the
main capacitor after passing the ZnO block. Shortly, the
downward current was passing through all brass rods, the
upward current was through the middle brass rod passing the
ZnO block, Rogowski Coil, and the main capacitor. By this
method, the construction inductance is suppressed to a very
small value as small as 3.5 H and its resistance value was as
small as 0,56 ohms.
Both two circular brass plates, each of which has 2
meter diameter, are connected by 8 brass rods 1.6 meter
length and 9 mm diameter of each, using bolts and screws. In
addition, 32 copper wires having 3 mm diameter were
connected to the two brass circular plates.
Spheresphere
Elektrodes

inductance
Rf

Driving
system

controlled to determine their trigger times required. The

duration between two consecutive trigger times can be set as
small as 1 mS up or as big as 1000 mS. Here is the simple
diagram of the mechanically system trigger, Fig.2.
Fig.3 is the 5 concecutive impulse current waveshapes
produced by the impulse current generator shown in Fig.2.
current

Fig 3 Five impulse current produced.

A current shunt, Fig.4, is the most important device
made for measuring the impulse current flowing through the
ZnO block. Basically, it consists of two parts, the outer part
is a hollow copper cylindrical tube, for downward current
flow and the upward current flows through the inner part.
The inner part consists of two circular metal plates connected
by a 31 thin Ni-Chrom wires , each of which of 0,2 mm
diameter (36 /m) and 7 cm length, making the total current
shunt resistance value of 0.029 ohm
Menuju ke pembangkit arus tinngi impuls

a)
Keterangan :
Metal (Cu)
Isolasi
Arah arus

b)

Transfer data opto

elektronis
Osiloscope
(Pengukur tegangan)

Fig.4 Current shunt

By triggering the 5 multiple impulse generator, Fig.1, and
seeing the effect of one impulse in more detail, it is clear that
the output voltage of the voltage devider1: Vovd1, for
calculating the ZnO current, recorded by the compuscope has
the same starting time with the output voltage of the voltage
devider2 : Vovd2, for the ZnO voltage determination, which
also was recorded using the same device. Both curves are
shown in Fig.3 below.

ZnO

Cs

Fig.2 Simple 5 multiple impulse current generator

In order to produce 5 high speed ( the impulse front
time is about 10 S) and high magnitude (10000 Amper
peak) multiple impulse currents, the main capacitor, Fig.1,
was replaced by 5 main capacitors completed with their own
trigger circuits, that can be mechanically or computerly
ISSN: 2085-6350

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

71

The ZnO voltage, VZnO, can be determined by :

VZnO = Vovd2 * VD2ratio Vcs . 4)

vi

V1
V3

time

dt

Fig. 3 Trapezoidal rule of integration

method for calculating area

4. Result and discussion

The current shunt voltage, Vcs, was obtained using :

Vcs

Vi
curve

V2

= Vovd1 * VD1ratio 2)

Arus Zno 0.6k30.6.5 Teg Kap 10 kVdc

1500
1000
500

323

181

39,1

-103

98

240

-44

299

157

-500

15,2

0
216

The matter of determining is actually not as simple as is

suggested by Eq.1) because both c and v-i characteristic are
temperature dependent. Other thing that should be
considered is that the energy is not absorbed linearly
throughout the block.
In this work, five multiple impulse voltage values, the
Vovd2 : the output voltage of the voltage devider2 for the
ZnO voltage and the Vovd1 : the output voltage of the
voltage devider1 for current shunt voltage, were recorded by
the compuscope together with their related times. It is
important to note that the compuscope storing data should be
made in ASCI modes.Next, the ASCI data should be
modified where point showing decimal values should be
replaced by comma in order to be abele to be properly
processed by Excel software. The Excel software in the
following step was used to access the ASCI data and to
process them. Due to the same starting time between the two
voltages recorded by the compuscope, Fig.3, then the
impulse energy was easily calculated by applying some
equations as follows :

Impulse Energy =Area1 + Area2 + .+ AreaN 7)

-127

where :
M = mass (kg)
c = specific heat capacity (J/(kg C))
v, i = associated voltage and current

-68

vidt 1)

The impulse energy absorbed by ZnO block is the

summation of all trapezoid under the vi curve, or if there are
N number of trapezoids then :

74,1

Mc=

Area 1 = (v1+v2)*dt/2
.. 5)
Area 2 = (v2+v3)*dt/2 .. 6)

275

The passage of an impulse current through a resistive

object results in energy absorption and hence temperature
rise[10],[11] For short duration impulses, heat loss can be
neglected for times less than 1 second, and hence the average
temperature rise may be determined from :

The area of one trapezoid of Fig 3 :

133

3. Methode of multiple impulse current

energy calculation

The power VZnO * IZnO can be easily calculated, after

which the instantaneous power can be plotted.. Futher, the
accumulated energy, after the implementation of the 5
multiple impulse current, absorbed by the ZnO block could
be determined by trapezoidal rule of integration [6] as shown
as :

-8,7

Fig. 3 ZnO current (upper trace) and ZnO voltage

(lower trace), starting at the same time.
The setting of oscilloscope:
50 S/div, and 1 V/div

-150

Starting point of the

impulse

where :
VD1ratio = 50
VD2ratio = 900

Arus ZnO, Amper

impulse

-1000
Waktu, mikrodetik

a)

The current flowing through the ZnO block, IZnO, was

calculated using :
IZnO = Vcx/Rcs .... 3)

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

72

Proceedings of CITEE, August 4, 2009

absorbed by the ZnO block. The energy of first impulse

should be added to the energy due to the other 4 impulses.
The impulse energy will only exist as long as the current is
not zero. During the zero current, there is no addition in the
energy, see Fig 4 c).

Tegangan ZnO 0.6k30.6.5 Teg Kap 10 kVdct

2000

323

181

39,1

-103

98

240

-44

299

157

15,2

216

-127

-68

74,1

275

133

-8,7

0
-2000

-150

Tegangan ZnO, volt

4000

-4000

5. Conclusion

-6000
Waktu, mikrodetik

A method of the energy calculation of impulse energy

absorbed by a ZnO block have been developed using ASCI
mode stored data records in a compuscope and then
processed by EXCEL and finally some plots produced. The 5
impulse current energy calculation and absorbed by ZnO
block was less than 50 J/cm3. by impul current peak of 600
A due to the DC voltage charging capacitor of 10 kV.

b)

50
40
30
20

6. References

10

[1]

217

-54

81,6

295

159

23,6

-112

237

101

-34

314

179

43,1

-92

256

-15

121

0
-150

Energi ZnO, Joule/cm3

Energi ZnO 0.6k30.6.5 Teg Kap 10 kVdc

Waktu, mikrodetik

c)
Fig. 4 The ZnO current, ZnO voltage and
energy absorbed of Zno block due to
impulse currents produced by 10 kVdc
charging the main capacitor of the
multiple impulse current generator
a) ZnO current,
b) ZnO voltage
c) ZnO accumulated energy absorbed
The compuscope cannot displayed the 5 impulse waveshapes
together in the same screen due to the maximum data allowed
to be displayed. This is the result from the fact that one
impulse will last about 100 S, and the time between
consecutive impulse is about 31 mS. Futher, the time step
used in the compuscope data recording is 0.1 S. So if the 5
impulses should be displayed on the same screen for about
155 ms it needs 1.550.000 data. This is to much for the
compuscope. Therefore, the compuscope can only record the
impulse one by one. Consequently, every impulse has similar
value of starting time and similar value of ending time or in
other words, every impulse is treated similarly by the
compuscope.. This limitation results as can be seen in Fig 4.
a), b), and c) especially in their horizontal axis, the time seen
at every starting time of each impulse has similar value, and
so be their ending times. However, this limitation doesnot
make any big different in calculating the impulse energy

ISSN: 2085-6350

International Standar, IEC 60099-4, 2006, Metal Oxide Surge

Arresters without Gaps for A.C. System. P.37
[2] Darveniza M., Roby D., and Tumma L.R., Laboratory and Analytical
Studies of the Effects of Multipulse Lightning Current on Metal Oxide
Araresters, IEEE Transactions on Power Delivery, Vol.9, No.2, April
1994. pp.764 - 771
[3] Lat M.V., Analytical Method for Performance Prediction of Metal
Oxide Arresters, IEEE Transactions on Power Apparatus and Systems,
Vol.PAS.=-104, No.10, October 1985 pp. 2665-2674
[4] Haddath, H., and Warne. D. Advances in High Voltage Engineering,
IEE Power and Energy Series 40, the Institution of Electrical
Engineering, MPBooks Limited, Bormin, Cornwall, 2004, p 229.
[5] Haryono T., K.T.Sirait, Tumiran, Hamzah Berahim, The Design of a
10000 Amper Impulse Current Generator Using Low Voltage
Components, High Voltage Engineering, Vol.34, No.12, Desember
2008, High Voltage Engineering Press, Wuhan 430074, China,
pp.2726-2731
[6] Haryono T., K.T.Sirait, Tumiran, Hamzah Berahim., The Computer
Simulation for High Amplitude Impulse Current Design, 2008
International Conference on High Voltage Engineering and
Applications, November 2008, Chongqing, China, pp.207-210.
[7] Haryono T., The Design of A High Amplitude Impulse Current
Generator, International Conference on High Voltage Engineering
and Applications, November 2008, Chongqing, China, pp.339-343.
[8] Haryono T., Sirait K.T., Tumiran, Hamzah Berahim, Current Shunt for
Large Impulse Current, ACED 2008 the 14th Asian Conference on
Electrical Discharge, Nov.23-25, 2008 Bumi Sangkuriang, Bandung,
Indonesia, pp.231-234
[9] Haryono T., Sirait K.T., Tumiran, Hamzah Berahim, Rogowski Coil for
the Measurement of High Impulse Current, ACED 2008 the 14th Asian
Conference on Electrical Discharge, Nov.23-25, 2008 Bumi
Sangkuriang, Bandung, Indonesia, pp.204-206
[10] Darveniza, M., Roby, D., and Tumma, L.R., 1994. Laboratory and
Analyitical Studies of the Effects of Multipulse Lightning Current on
Metal Oxide Arresters, IEEE Transactions on Power Delivery, Vol.9,
No.2, April 1994, pp 764 771
[11] Darveniza M., Mercer D.R., 1993, Laboratories Studies of the Effects
of Multipulse Lightning Currents on Distribution Surge Arresters,
IEEE Transactions on Power Delivery, Vol.8, No.3, July, 1993.

Conference on Information Technology and Electrical Engineering (CITEE)

73

Proceedings of CITEE, August 4, 2009

STUDY OF RETURN VOLTAGE TRANSIENT IN LOW VOLTAGE ZnO

ARRESTER TYPE OBO BETTERMANN V-20 C/1
Nurcahyanto1 , T. Haryono2, Suharyanto.3
1,2,3
Electrical Engineering Department,University of Gadjah Mada Yogyakarta
Jalan Grafika No 2, Kampus UGM Yogyakarta 55281,Indonesia, phone :0811250611
Website: http://te.ugm.ac.id/
Abstract : Arrester is an equipment to protect a power
system equipment from overvoltage. So, an experiment of
the arrester must be done accurately and appropriately
based on international standard. Arrester allows high
current flowing the ground. If the current flowing through
the arrester is too large, arrester may be damaged. From
all of the techniques to investigate the degradation of
arrester material, Return Voltage Measurement (RVM)
have been done excessively. Many people use RVM method
because this method has a high sensitivity to detect the
degradation of the arrester. Commonly, experiment study of
the arrester involves the use of a single or multiple current
impulse striking the arrester. The goal of this experiment is
to model the arrester in return voltage condition which can
be seen as its phenomena a return voltage having been
known. The experiment of RVM can be continued
investigating the influence of lightning voltage impulse
strike to the arrester. RVM was done to an arrester before
and after the impulse implementation of impulse strike to it.
Its clear that the highest DC value of return voltage
obtained was 674 mV resulted from charging voltage of 200
volt after experiment and recorded in 5 second of charging
time. The value is not much different from the arrester
before experiment which has DC voltage 512 mV in the
same treatment.
Keywords : ZnO arrester model , return voltage, lightning
impulse voltage
1. INTRODUCTION
The principle way to identify a return voltage is the use
of equipment which is able to charge, switch, and measure
the return voltage accurately. Commonly, the experiment
uses the equipment that is called Tettex RVM 5462[1]. Tettex
RVM 5462 consists of a DC source, a serial switch, a
parallel switch, and an electrometer. Tettex RVM 5462 was
able to do a switching process in micro or milli second.
However, in this research Tettex RVM 5462 was not
available, so the research was done with the other devices
that were able to represent Tettex RVM 5462 circuit
although it has lower accuracy than Tettex RVM 5462. The
measurement involved a digital voltmeter. In this method,
all of the devices had been set and appropriately for the
research and then were done the return voltage records.

experiment was OBO BETTERMANN V-20 C/1. Metal

oxide arrester can be modeled by Rp, Cp, Rg, and Cg[6].

Rp
Cg

Rg

Cp

Fig.1 Low Voltage ZnO arrester model

with :
Cg = geometric capacitance
Rg = geometric resistance
Cp = polarisation capacitance
Rp = polarisation resistance
The arrester model involves polarisation resistance
and polarisation capasitance. In this polaritation, the arrester
is represented as of electron and hole that move in the
oponent polarity. Electrons move in the positive polarity
and holes move in the negative polarity. This movement
occures when a DC source is applied and then is opened. If
the circuit is shorted and then voltage shows aproximately
to 0 V, and it is believed the arrester being completely
discharged. However there is still some voltage between
arrester terminals. This voltage is due to the polarization of
insulation. The movement of electron is limited by
polarisation resistans of ZnO material. Geometric
polarisation is a value of the arrester that can be represented
as a paralel connection between R and C. A geometric
capasitor is an ideal model of capasitor. On the other hand,
geometric resistance characterize the resistance of ZnO
arrester insulation. Basically the research circuit of the
return voltage can be shown by the picture below:

S1

2.

MODELING & EXPERIMENTAL SET UP

The object of this experiment was metal oxide arrester

(ZnO). ZnO arrester consists of 90% zinc oxide and 10 %
others metals[4,5]. The arrester type used throughout this

DC

S2

Arrester

V voltmeter

Conference on Information Technology and Electrical Engineering (CITEE)

Fig.2 Basic Research Scheme

ISSN: 2085-6350

Proceedings of CITEE, August 4, 2009

74

U
Uc
next
cycle

UDC

Fig.3 Design Circuit of Research

tc

td

trelaxation

Fig.7 Fourth Step

The measurement procedure consists of four test steps[2,3].

1. Charge the select object for a pre-select time (tc) with
a DC voltage (Uc) (switch S1 is closed), which is
much lower than the rated voltage.
2. Discharge the test object for a short period of time
which is normally half of charging time, td=1/2 tc.
(switch S1 is open and switch S2 is closed)
3. Measure the open circuit voltage (Switch S2 opens and
at the capacitor electrodes the voltmeter registers
voltage produced by the residual capacitor charge, this
is called DC return voltage).
4. After recording the return voltage, the switch S2 is
closed and the test object is short circuited.

U
Uc

UDC
tc

td

U
Fig.8 Return Voltage Graphics

Uc
From the graph Fig.8,it can be seen that the research was
done by charging the arrester using a DC voltage (Uc) for a
pre-selected time (tc), and then it is discharge during for a
short period of time (normally half of charging time, td=1/2
tc, switch S1 is opened and switch S2 is closed). An then
after the switch S2 was opened, the capacitor produced
some value of voltage that was called the residual return
voltage. The return voltage occurred because arresters
characteristics that still had stored energy charged. The
value of return voltage was measured by voltmeter. The
transient phenomena of return voltage was recorded only
the DC measurement voltage and the value decreased
according to the picture above.

tc
Fig.4 First Step

U
Uc

tc

td

t
3.

Fig.5 Second Step

U
Uc
UDC
tc

td

Return Voltage Measurement Experiment

applied to OBO BETTERMANN V-20 C/1
Arrester before and after applied by Lightning
Voltage Impulse

In this experiment, it used a voltage impulse to strike an

arrester. There were two major experiments done before and
after experiment. It means the experiment of return voltage
was done before the arrester was stroken by lightning
impulse voltage and after the arrester was stroken by
lightning impulse voltage. The lightning impulse was
divided into 3 three stages:

Fig.6 Third Step

ISSN: 2085-6350

Conference on Information Technology and Electrical Engineering (CITEE)

75

Proceedings of CITEE, August 4, 2009

Stage I

:arrester was applied once the lightning voltage

impulse of 4,36 kV.
Stage II
: arrester is applied three times by lightning
impulse voltage (4,87 kV,6,04 kV and 6,77
kV)
Stage III : arrester is applied four times by lightning impulse
voltage (7,68 kV, 8,49 kV, 9,37 kV and 10,31
kV).
Table 1.1 Values of DC return voltage at the 200 Volt
Charging Level before Experiment

DCVoltage
200Volt
200volt
200volt
200Volt
200Volt
200Volt
200Volt

ChargingTime
(tc)
5second
10second
20second
40second
60second
80second
100second

DCReturn
Voltage(mili
Volt)
512
275
147
74.5
53
41.3
37.1

Fig.10 Graphic Return Voltages Comparison from

Experiment
Fig.10 show the experiment results in terms
different values of return voltages before and after
experiment. After the experiment, a higher return voltage in
some of research value (different time record) produced,
however the return voltage increase was not so significant.
Other experiment resoluted in other value if it used current
impulse[2]. By the use of impulse current, the value of return
voltage was tend to decrease[2], It can be concluded that if
the arrester applied impulse voltage is not enough to destroy
the material of arrester, so the arrester still has a good
condition. in the other words, it was clear that there was no
difference because the difference only a few mili Volt, so it
could be concluded that the characteristics of arrester was
still the same and didnt change of characteristics, before
and after experiment, and the voltage impulse didnt have
significantly effect to damage arrester.

4.
Fig.9 DC Return Voltage at 200 Volt Charging Level

Table 1.2 Return Voltage after Experiment

DCVoltage
(miliVolt)

Charging
Time

200volt
200volt
200volt
200volt
200volt
200volt
200volt
200volt

5second
10second
20second
40second
60second
80second
100second
120second

DCRetrurnVoltageafter
Experiment(miliVolt)
I
II
III
497
526
674
339
291
403
221
249
197
124
95.9
108.9
108.1
88.2
78.7
101.3
77.1
77.7
100
79
55.8
86.7
57.6
40.5

CONCLUSSION

In this research have some conclusion :

1. Arrester with the ZnO material have a capasitor
characteristics, so it can store a charge.
2. Return Voltage Measurement is one method that is
available for determining arrester characteristic.
3. The technique of return voltage measurements is
used to indicate any lightning strike experiment
given to arrester :
Arrester was made from specific material, in
this case made from zinc oxide which have a
high resistance.
Arrester must be through a high current not
only high voltage.
4. Stroken high voltage arrester by will not probably
make any damage to arrester material because the
current flowing to the arrester is not too high.
5. The return voltage is not endanger the system
because the voltage value is not too high.

Conference on Information Technology and Electrical Engineering (CITEE)

ISSN: 2085-6350

Proceedings of CITEE, August 4, 2009

76

5.
[1]

[2]

[3]

REFERENCES
Zulkurnain,Abdul-Malek, Novizon, Aulia,2005 ,
Return Voltage Measurement to Investigate the
Degradation of Zinc Oxide Varistor.
K.P. Mardira, T.K Saha, R.A. Sutton, 2005,
Investigation of Diagnostic Techniques for Metal
Oxide Surge Arresters.
Jaroslaw Gielniak, Marek Ossowski, 2005, Dielectric
Response of Oil-Paper Insulation Systems of Large
Moisture and Temperature Inhomogeneity.

ISSN: 2085-6350

[4]

[5]

[6]

Anonim,October 1999 ,Zinc Oxide Arrester Design

and Characteristics, Hubbel Power Systems Inc,
USA
Lalu Suryodarminto, 2007 ,Pengaruh Suhu Pada
Tegangan Discharge Blok Zno. Tugas Akhir Jurusan
Teknik Elektro Universitas Gadjah Mada.
http://www.haefely.com/pdf/LL_RVM5462.pdf

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

77

The Design of Digital Overcurrent Relay

with IEC 60255 Time Curve Characteristic
Based on ATmega16 Microcontroller
Agni Sinatria Putra1, Tiyono2, Astria Nur Irfansyah3
1

Student of Electrical Engineering Department, Gadjah Mada University

Lecturer of Electrical Engineering Department, Gadjah Mada University
3
Lecturer of Electrical Engineering Department, Gadjah Mada University
2

Abstract
The use of digital technology ensures the accuracy of protective relay operation, especially in the
protection algorithm and operating time. Digital technology is employed through a lightweight
and compact microcontroller, an ATMega16, which is expected to have long operating life and
simple maintenance. The ATmega16 has shown satisfactory results in performing computations for
the protection algorithm, especially in over-current protection. To do current sensing, the current
flow is read through the Hall effect sensor. The function of time-current curve characteristics
works according to the IEC 60255 standard and to an inverse function that can be set based on
user requirements. The fastest relay operating speed is limited with consideration to reduce the
error of decision-making in transient condition, and the response time of signal conditioner
circuit. The peak-value reading method is found to provide the most rapid response, with the
update every 10 milliseconds. Tests with the dynamic and static load indicate that the protective
relay designed is immune to the transient phenomena such as DC offset and the motor starting
currents.
Keyword: Protective Relay, Over-current, ATmega16 microcontroller, Time-Current Curve
1. Introduction
The Institute of Electrical and Electronic
Engineers (IEEE) defines a relay as an electric
device that is designed to respond to input conditions
in a prescribed manner and, after specified conditions
are met, to cause contact operation or similar abrupt
change in associated electric control circuits. A
note adds: Inputs are usually electric, but may be
mechanical, thermal, or other quantities or a
combination of quantities. Limit switches and similar
simple devices are not relays (IEEE C37.90).
The IEEE defines protective relay as a relay
whose function is to detect defective lines or
apparatus or other power system conditions of an
abnormal or dangerous nature and to initiate
appropriate control circuit action (IEEE 100).
The function of protective relaying is to cause the
prompt removal from service of any element of a
power system when it suffers a short circuit, or when
it starts to operate in any abnormal manner that might
cause damage or otherwise interfere with the
effective operation of the rest of the system. A
secondary function of protective relaying is to
provide indication of the location and type of failure.
The fundamental objective of system protection is
to provide isolation of a problem area in the power
system quickly, so that the shock to the rest of the
system is minimized and as much as possible is left
intact. Within this context, there are five basic facets
of protective relay application:

Reliability: assurance that the protection

will perform correctly.
2. Selectivity,
related
to
protection
coordination, to keep the maximum
continuity of service with minimum system
disconnection.
3. Speed of operation, minimum fault duration
and consequent equipment damage and
system instability.
4. Simplicity, minimum protective equipment
and associated circuitry to achieve the
protection objectives.
5. Economics, maximum protection at minimal
total cost.
An overcurrent relay is a relay with an inverse
time characteristic, (inverse means that the relay
operates faster as current increases) that operates
when input current exceeds a predetermined value,
called as pickup value. On the time-curve
characteristic, the value of current stated as the
multiple of pickup. The time-curves characteristic
have a several operating time scale, known as Time
Multiplier Setting (TMS). The relay contact operates
when the duration of the input current exceeds the
operating time.
Time-current curves are designated as inverse,
moderately inverse, very inverse, and extremely
inverse. A definite time relay is one that operates at
the same time, at all current magnitudes above the
pickup set point. The degree of inverseness is an
indication of the change between the relay pickup at

Conference on Information Technology and Electrical Engineering (CITEE)

1.

SSN: 2085-6350

78

Proceedings of CITEE, August 4, 2009

1.5 times pickup and 10 times pickup. Operate time,

when the multiple is greater than 1, but less than
1.25, may not be shown on time-current curves as
repeatable operation may not be guaranteed. IEC
60255 defines a number of standard characteristics of
IDMT relay as follows:
Table 1 IEC 60255 characteristic equation

  

Standard Inverse (SI)

Very Inverse (VI)

  

Extremely Inverse (EI)

  

Definite Time (DT)

0.14

.


1

13.5

(1)

(2)



 1

80



 1


 , ! 

(3)

the fundamental frequency (50/60 Hz). The AC

signal reading algorithm is applied to obtain the
value of the component magnitude. The data is used
by microcontroller to do the protection algorithm.
Microcontroller will monitor the abnormal condition
based on the algorithm. This algorithm is such a kind
of equations that designed by protection engineer and
protective relay manufacture. If a fault detected, the
contact will operate and trip the associated CB.
2. Research Methodology
The one phase protective relay designed is a
laboratory scale one with specifications:
a. Based on ATmega16 Atmel microcontroller
family.
b. Using a linear and accurate Hall Effect current
sensor that can sense the current up to 20 A
peak with 2,1 kV terminal insulation.

(4)

IEC 60255 time-curve characteristic illustrated in

Fig. 1 as follows:

c. Operate based on International Electrotechnical

Commission (IEC) 60255 standard and one
user-defined time-current curve.
d. Have a current and a operating time set point on
the definite setting.
e. TMS scale from 0,05 up to 1, with interval in
0,05.

Operating
time (seconds)
Waktu Kerja (detik)

f. Using a LCD 32 character display and 4 x 4

matrix keypad user interface for relay
parameter setting input.
System block diagram shown in Fig. 2 consists
of:
Standard Inverse (SI)

1.

Current sensor

2.

Signal Conditioner circuit

Definite time (DT); t=1

a.

amplifying and null offsetting circuit

b.

Band Pass Filter

c.

Precision rectifier

d.

Signal scaling circuit

Very Inverse (VI)

Extremely Inverse (EI)

Kelipatan Arus Pickup

Multiple of pickup

3.

Polarity detector

4.

Primary microcontroller

5.

Auxiliary microcontroller

6.

CB tripping circuit

7.

32 character LCD display

8.

Matrix 4 x 4 keypad

Fig. 1 IEC 60255 characteristics on TMS=1

The digital relay require analog to digital
converter (ADC) that sampling the signal from the
current transformer and voltage transformer into data
in digital form. On several digital relay, the data
sample from the ADC can be saved for oscillosgraf
purpose, but for use in protection programming
algorithm, only the fundamental frequency
component is required.
The data sample from the ADC is feed through
the band pass filter to remove the frequencies outside
SSN: 2085-6350

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

79

information to auxiliary microcontroller via

USART serial communication.
5.

Microcontroller compares the current magnitude

sensed with the pickup value of relay setting. If
the value read exceed the pickup setting, then the
microcontroller will wait for 100 ms long to
update the current reading. This done to
anticipate reading error caused by DC offset
value on fault current. After 100 ms, DC offset is
expected to become further smaller and the
current reading is valid. The other purpose is to
make the relay immune against magnetizing
inrush current that happened instantaneous as an
inductive load energized. After this second
reading, if the current value sensed still exceed
the determined setting, then microcontroller will
run the time delay due to time-curve
characteristic settled.

6.

While running the time delay, microcontroller

will recheck the current magnitude by ADC
program routine, every 1000 ms, just in case the
time delay is more than 10 seconds. The recheck
operation is intended to avoid the fals
falsee decision
in running the time delay if fault was occurred
temporary.

7.

If the fault is temporary occurred, then go back

to point 4.

8.

After time delay is done, microcontroller will

once again sense the current magnitude. This
done to ensure whether it is not temporary fault.
If the fault is temporary occurred, then go back
to point 4. Or, if the sensed current is in the
condition of normal operation that resemble
then
n the CB
fault, such as inrush current, the
tripping command canceled.

9.

command
mmand via one of
Relay gives CB tripping co
port that connected to external circuit to trip the
CB.

Fig. 2 System Block Diagram

System work:
1. The Hall effect sensor sense current on the
line. The result
result is analog signal that reflect
the sensed current in prescribed scale.
2.

3.

1.

The function of signal conditioner is to

provide treatment to the particular signal
input through the parts as follows:
a.

The amplifying and null offsetting

block will remove the offset value that
came from the current sensor and
provide gain in predetermined level to
increase the signal to noise ratio.

b.

Band Pass Filter removes frequencies

outside the 50 Hz fundamental value.

c.

Precision rectifier will rectify sine wave

signal from band pass filter to be full
wave dc signal.

Main microcontroller will read the peak

value from the full wave signal for every
half cycle of sine wave (10 ms). This
reading method results
results in a fast response
time of current magnitude
magnitude change compared
with other methods that was previously
attempted.

4.

The auxiliary microcontroller will provide

the information of relay setting parameter,
inputted by user, which is needed by main
microcontroller for the operation.

5.

The CB tripping circuit acts as an interface

between main microcontroller and CB.

The microcontroller algorithm:

First initialization.
initialization.

2.

Microcontroller call the relay setting values

saved on EEPROM.

3.

Microcontroller applies the relay setting.

4.

Microcontroller read the current magnitude

through the ADC, and send the current

10. Microcontroller sends the fault information to

the auxiliary microcontroller through USART
serial communication, and activates alarm.
11. Microcontroller waits the reset
interruption to restart the routine.

button

3. Result and Evaluation

3.1. Signal Conditioner Circuit Testing
The signal conditioner circuit give
givess a good
linearity and expected output. DC offset value of the
input signal have been removed and the signal is
rectified to become
become full wave DC signal with double
gain. Those signals shown in
in Fig. 3 below.

Conference on Information Technology and Electrical Engineering (CITEE)

SSN: 2085-6350

80

Proceedings of CITEE, August 4, 2009

Standard Inverse characteristic testing shown on

Fig. 5. The recorded error is 1,78 %. The thin red line
curve is the curve computed theoretically. The
designed relay is able to complete Standard Inverse
equation for all TMS value. Red shaded area is place
where the relay has 8 % error.

magnitude
input

output
Standard Inverse Time-Current Curve with Definite Time

Fig. 3 The input and output signal of the signal

conditioner
The implemented polarity detector is able to
provide low logic level for positive polarity and high
logic level for negative polarity of sine wave input.
Fig. 4 shows the output signal of the polarity detector
for a sine wave input.

Operating time (seconds)

time

magnitude
High Level (-)
Multiple of pickup

Input

Fig. 6 shows the addition of definite setting. The

operating time will directly set on the current value
greater than the value of definite setting
satisfactorily. This result is also valid for the other
characteristics curve testing in Fig. 8 and Fig. 10.

time

Fig. 4 The result of polarity detector

3.2. Relay Characteristic Testing
This test is conducted by generating a sinusoidal
signal using AFG that resembling the output of Hall
effect current sensor. The test is meant to know that
the relay has been working according to the desired
characteristics. The pickup value setting is 2 A and
the red shaded area is the value of multiple pickup
current 1 - 1.3
3.2.1. Standard Inverse

Operating time (seconds)

Standard Inverse Time-Current Curve

3.2.2. Very Inverse

Very Inverse Time-Current Curve

Operating time (seconds)

Low Level
(+)

Fig. 6 The result of Standard Inverse with definite

time mode

Multiple of pickup

Fig. 7 The result for Very Inverse characteristic

Fig. 7 above is the result of Very Inverse
characteristic testing which has recorded error 1,8 %.
The designed relay is able to complete Very Inverse
equation for all TMS value. Red shaded area is place
where the relay has 10 % error.

Multiple of pickup

Fig. 5 The result for Standard Inverse characteristic

SSN: 2085-6350

Conference on Information Technology and Electrical Engineering (CITEE)

Proceedings of CITEE, August 4, 2009

Very Inverse Time-Current Curve with Definite Time

81

3.2.4 User-defined Characteristic

Operating time (seconds)

Curve equation:

  

"



 1

which are K and n can be selected as needed.

Operating time (seconds)

User-Defined Time-Current Curve

Multiple of pickup

Fig. 8 Very Inverse with definite time mode

3.2.3. Extremely Inverse

Operating time (seconds)

Extremely Inverse Time-Current Curve

Multiple of pickup

Fig. 11 The result of custom invers characteristic

Multiple of pickup

Fig. 9 The result of Extremely Inverse characteristic

The recorded error is 2,1 %. Fig. 9 plotted the

result of Extremely Inverse characteristic testing. The
thin red line curve is the curve computed
theoretically. The designed relay is able to complete
Extremely Inverse equation for all TMS value. Red
shaded area is place where the relay has 10 % error.
This is an reasonable area for all overcurrent relay to
have larger error, un-guaranteed operation area.

Fig. 11 shows that the custom Inverse can be done

with recorded error is 4,5 %. The thin red line curve
is the curve computed theoretically. Red shaded area
is place where the relay has 10 % error. The greater
the constant K will result in the longer the relay
operate, while the greater the value of n will result in
the form of an increasingly steep curve.

3.3. Test Against DC Offset

The next following tests is using the electrical
circuit consist of a current source and loads such as
resistive load and electric motor. Sensed current are
shown in the osillosgraf recording.

Operating time (seconds)

Extremely Inverse Time-Current Curve with Definite Time

Fig. 12 Test againts dc offset

Multiple of pickup

Fig. 10 The result for Extremely Inverse characteristic

with definite mode

DC offset testing in Fig. 12 indicates that the

relay remains to read the current magnitude after dc
offset decayed, thus the error in reading can be
avoided.

Conference on Information Technology and Electrical Engineering (CITEE)

SSN: 2085-6350

82

Proceedings of CITEE, August 4, 2009

2. With consideration to reduce the error in the

decision making on transient condition and
the response time of signal conditioner, the
fastest relay speed is 100 ms.

3.4. Test Against Temporary Fault

3. Protective relay designed perform functions

of
International
Electrotechnical
Commission (IEC) 60255 standard timecurrent curve. Plus an inverse function with
the constants regulator that can be set
according to the curve desired.
Fig. 13 Test againts temporary fault
Temporary interruption does not cause relay to
give tripping signal. Temporary interruption occurs
in a short time less than relay operating time, shown
in Fig. 13.
3.5. Test with Static Load

4. The method of peak value reading is found

to give the best response to the major
changes of current flow with the update rate
in reading every 10 ms.
5. Tests with the dynamic and static load
indicate that the protective relay designed
immunes to the transient phenomena such as
DC offset and the starting motor currents.

5. Reference

Fig. 14 Test with static load

Relay gives tripping signal during a phase to the
ground fault occurs and relay immune to the dc
offset. Fig. 14 illustrates the sine wave of the current.
3.6. Test with Dynamic Load

Fig. 15 Test with dynamic load

Test with dynamic load run successfully in Fig.

15. At the time when motor starts, relay does not give
tripping signal, but when the single phase to ground
fault occurs on the line.

[1] Atmel,
ATmega16
Datasheet.
http://www.atmel.com/literature. Atmel. 2005.
[2] ALSTOM T & D Energy Automation &
Information,
Network
Protection
&
Automation Guide, Alstom. France, 2002.
[3] Blackburn, J. Lewis, Domin, Thomas J.,
Protective Relaying Principles and Application
3rd Edition, CRC Press, Taylor & Francis
Group, LLC, Boca Raton, 2006.
[4] Elmore, Walter A., Protective Relaying:
Theory and Application 2nd Edition, Marcel
Dekker, Inc., New York. 2003.
[6] GE Power Management, GE Product Catalog,
General Electric, Co., USA. 1999.
[7] Hase, Yoshihide, Handbook of Power System
Engineering, John Wiley & Sons, Ltd.,
England. 2007.
[8] Mason, C. Russel, The Art and Science of
Protective Relaying, John Wiley & Sons Inc.,
USA. 1956.
[9] Mavino, Albert Paul., Electronic Principles 2nd
Edition, McGraw-Hill, Inc., New York. 1979.
[10]Ravindranath, B., Chander, M., Jha, C.S.,
Power System Protection and Switchgear
Wiley Eastern Limited, New Delhi. 1976.
[11] Sleva, Anthony F., Protective Relay
Principles CRC Press, Taylor & Francis
Group, LLC, Boca Raton. 2009.
[12] Tocci, R.J., Widmer, N.S., Digital Systems
Principles and Application Prentice Hall
International, New Jersey. 1977.

4. Summary and Conclusions

1. ATmega16 is a computation tool that work
quickly and accurately. Very well to be used
to do the algorithm protection on a digital
protective relay.
SSN: 2085-6350

Conference on Information Technology and Electrical Engineering (CITEE)