Professional Documents
Culture Documents
FOREWORDS
Dean of Faculty of Engineering, University of Indonesia
The Quality in Research (QIR) Conference is the annual event organized by the Faculty of
Engineering, University of Indonesia. Since started in 1998, it has become an excellent forum of
discussion for all researchers from research institutions and universities all over the nation of
Indonesia. The 1
st
and 6
th
QIR Conferences had been successfully organized as the high quality
national conferences, and starting from the 7
th
QIR conference, has been organized to invite
international research papers.
The 10
th
Quality in Research International Conference having a theme of Research for future better
life is to provide an international forum for exchange of the knowledge, information, experience and
result as well as the review of progress and discussion on the state of the art and the future trend
various issues and the developments in the multi-fields of scientific and technology. The main
purposes of this conference are to provide a forum for free discussion of new ideas, development and
applications, including techniques and methods to stimulate and inspire pioneering work, to provide
opportunities for students and young engineers to meet their experienced peer and to provide a
meeting that will enforce progress, stimulate growth and advance the state of knowledge in the multi-
fields of science and technology.
We would like to express our heartiest to thank to all authors and participants for their active
participations in the 10
th
on Quality in Research (QIR) International Conference 2007, and also to all
the paper reviewers, member of the technical committees, and member of the organizing committees,
for their support to the success of this conference. Last but not the least; we would also like to invite
all participants to the next on Quality in Research (QIR) Conference.
Faculty of Engineering, University of Indonesia
Dean,
Prof. Ir. Rinaldy Dalimi, Ph.D
ii
FOREWORDS
Chairman of 10th International Conference on QIR 2007
The 10
th
Quality in Research International Conference will provide an international forum for
exchange the knowledge, information, experience and recent researches of various fields. With a
strong support and presentations from academic, industry and entrepreneurs, the conference will
provide an ideal platform to learn various fields and understand technological trends in the region.
The 10
th
Quality in Research (QIR) International Conference has a theme of Research for Better
Future Life being the third time to go internationally, has invited limited papers from other nations
such as Korea and Malaysia. The conference is organized in parallel sessions focusing on the 8 (eight)
research areas such that many researchers and peer groups may focus their discussion on the relevant
topics. All submitted papers had been reviewed by the technical committees and had been arranged
into 8 (eight) sub-themes according to the following fields:
Energy, Process and Environmental Engineering and Management: Energy and
environmental issues, combustion technology, fluids mechanics and thermal fluid machinery,
thermodynamics and heat transfer, geotechnical and environmental engineering, etc.
Industrial, Manufacturing, Material Engineering, and Management: Production
Engineering, Supply Chain Management, Innovation System, Maintenance System, Quality
Management System, Human Factors Engineering, Organizational System, Fabrication and
Industrial Automation, Manufacturing System: Control Management and Information
Technology, etc
Biomaterial, Biomedical Engineering and Biotechnology: Biomedical numerical modeling,
Biomaterial, Biosensor, Biocompatibility, Biomechanics, Biotechnology, Biomedical
Instrumentation, Biomedical Imaging
Design and Infrastructure Engineering and Management: Product design and
development, composite: Materials and applications, structural dynamics, mechanics of
materials, Construction Management, Public Infrastructures and Services, Structural
Engineering, etc
Special Session on Electronics Engineering
Information and Computation Technology
Sustainable Architecture
Nanomaterials and Nanotechnology: Nano structured material, Nanotechnology,
Nanocomposite, MEMS, Self Assembled Monolayer, Thin Film, etc
The main purposes of this conference are to provide a forum for free discussion of new ideas,
development and applications, including techniques and methods to stimulate and inspire pioneering
work, to provide Opportunities for students and young engineers to meet their experienced peer and to
provide opportunities for students and young engineers to meet their experienced peer and to provide
a meeting that will enforce progress, stimulate growth and advance the state of knowledge in the
multi-fields of science and technology.
Depok, 4 December 2007
The Organizing Committee,
Chairman,
iii
Gunawan Wibisono, Ph.D
Steering Committee
1. Prof. Dr-Ing. Axel Hunger, Universitaet, Duisburg-Essen, Germany
2. Prof. Dr. Carlo Morandi, Universida Degli Studi de Parma, Italy
3. Prof. Dr. Iwao Sasase, KEIO University, Japan
4. Prof. Kim Kyoo-ho, Yeungnam University, Korea
5. Prof. Dr. Ir. Irwan Katili, University of Indonesia
6. Prof. Dr. Ir. Bambang Suryawan, MT, University of Indonesia
7. Prof. Dr. Ir. Dadang Gunawan, M.Eng, University of Indonesia
8. Prof. Dr. Ir. Johny W Soedarsono, DEA, University of Indonesia
9. Prof. Ir. Gunawan Tjahjono, M.Arch, Ph.D, University of Indonesia
10. Prof. Dr. Ir. M. Nasikin, University of Indonesia
11. Isti Surjandari, Ph.D, University of Indonesia
12. Prof. Dr. Ir. Budi Susilo Soepandji, University of Indonesia
13. Prof. Dr. Ir. Sutanto Soehodho, University of Indonesia
14. Prof. Dr. Ir. Sulistyoweny Widanarko, Dilp. SE. MPH, University of Indonesia
15. Prof. Dr. Ir. I Made Kartika Dipl. Ing, University of Indonesia
16. Prof. Dr. Ir. Tresna P. Soemardi, University of Indonesia
17. Prof. Dr. Ir. Sardy, M.Eng, M.Sc, University of Indonesia
18. Prof. Dr. Ir. Bagio Budiardjo M.Sc, University of Indonesia
19. Prof. Dr. Ir. Djoko Hartanto, M.Sc, University of Indonesia
20. Prof. Dr. Ir. Eddy Siradj, M.Eng, University of Indonesia
21. Dr. Ir. Kemas Ridwan K, University of Indonesia
22. Prof. Dr. Widodo Wahyu P, DEA, University of Indonesia
23. Ir. Boy Nurtjahyo M.,MSIE, University of Indonesia
24. Dr. Ir. Dedi Prihadi DEA, University of Indonesia
25. Ir. Hendri D.S. Budiono, M.Eng, University of Indonesia
26. Dr. Ir. Sigit Pranowo Hadiwardoyo, DEA, University of Indonesia
27. Dr. Ir. Herr Soeryantono, University of Indonesia
2 28 8. . Prof. Rinaldy Dalimi, Ph.D, University of Indonesia
Chairman of the Conference
Ir. Gunawan Wibisono, M.Sc, Ph.D
Technical Committee
1. Ir. Gunawan Wibisono, M.Sc, Ph.D
2. Dr. Yosia Irwan
3. Dr. Engkos Kosasih
4. Purnomo Sidi P, Ph.D
5. Abdul Muis, Ph.D
6. Badrul Munir, Ph.D
7. Tania Surya Utami, MT
8. Ir. Beatrianis, M.Si
Foreword From The Dean of Faculty Engineering, UI
Foreword From Chairman of 10th International Conference on QIR 2007
The Committee of 10th International Conference on QIR 2007
Paper No. Title and Name of Author(s)
EPE-01 The Effect of Harmonics Current to Performance of The Over Current Relay
by: Syafrudin and Dahaman
EPE-02 Effect of The Harmonic Current Component to Active Power Losses on Power Transformer
by: Syafrudin and Dahaman I
EPE-03
Optimize placing of passive filter and reduce harmonic in power system, Case study at PT ISPAT
INDO
by: Iwa Garniwa, Budi S, Adrianto
EPE-04
Convergence Approximation Of ABR Formulation For Josephsons Tunneling in UO
2
Chain
Reaction at 45.7 MW Candu Nuclear Reactor
by: Moh. Hardiyanto
EPE-05 Energy Conservation on Steel Industri By Increasing Efficiency and Electric Power Quality
by: Iwa Garniwa, Budi S, Fauzan H J
EPE-06 Energy Conservation on Hotel Building (Case Study in Bandung)
by: Rudy S and Aji Nur W
EPE-07 Indonesia Energy Revolution Scenario and Projection using MESAP/Planet Simulation Model
by: Rinaldy D, Bayu I, and Sven Teske
EPE-08 Ocean wave power plant with air pressure
by: Massus Subekti and Vivian K L
EPE-09 The Time for Indonesia to Optimize Geothermal Energy
by: Massus Subekti
EPE-10
Incorporating energy Commodity Price Volatility in economic Cost of Supply Analysis for
Electricity Expansion Plan
by: Emil E. Dardak
EPE-11
Effects of Chimney Depth and Downjet Height in A Coal briquette Stove on the Co Emission and
Ignition Time
by: Dijan S., Yulianto S.N., and Dian N. Kusuma
EPE-12 Scale up Reactor for Bio-gasoline Production from Crude Palm Oil
by: Bambang HS., Anondho W., and M. Nasikin
EPE-13 Flame Lift-up on A Bunsen Burner; A Preliminary Study
by: C Prapti Mahandari and I Made Kartika D
EPE-14
Study of Bio-Diesel of Coconut and Corn Ethyl Ester Use By processing With The Processor
Series Type in Diesel Engine Performance
by: B. Sugiarto, Sanggul H. Siregar, and Yanuar C.
EPE-15 On The Measurement of Smoke Production rate of Tropical wood
by: Yulianto S. N.
EPE-16 Catalytic Ozonation of Endosulfan in Water With Activated carbon
by: Enjarlis, Setijo Bismo, Slamet, and Roekmijati WS
TABLE OF CONTENTS
EPE-17
Hydrocarbon Selectivity of Fisher Trosch Synthesis from syngs with ratio of H2 to Co=1.0over Co-
Fe/AI2O3 bimetalic catalysts: effect of Fe:Co ratio
by: Dewi Tristantini and Borje Gevert
EPE-18
The deactivation behaviour of HZSM-5 catalys at various temperature of acetone conversion to
mono-aromatic hydrocarbon
by: Setiadi, Prilly F J, and Toshinori K
EPE-19 CO2 Absorpation Through Hollow Fiber Membrane Gas-Liquid Contactors
by: Sutrasno K
EPE-20
Enzymmatic and acid hydrolysis of bagasse for ethanol production by simultanious sacharification
and fermentation
by: Samsuri, Gozan, Nasikin, and Prasetya
EPE-22 Biogrease using modified palm oil as base oil and thickener lithium soap
by: Sukirno, M. Nasikin, B. Heru, Rizqon F, Marius, Dizzi
EPE-21
A general Application of A Direct Method For Multivariable MPC Control Strategy in Chemical
Process
by: A. Ahmad, A. Wahid
EPE-23
Model for Oxidation and Combustion of Ethanol in Wide range of Pressures, Temperatures and
Equivalence Ratios
by: Yuswan Muharam and Soraya S
EPE-24 Water and air Circulation System in Trimaran Fishing Veseel Live fish holds
by: Sunaryo, Anita Muslih
EPE-25 Local Protein Structure Comparison
by: Ford Lumban Gaol and Belawati W
EPE-26
Antibacterial activity Analysis of Dellenia indiaca's Kernel Extract for Escherichia coli: A
Preliminary Study
by: Rita Arbianti, Tania S U, and Ifa P
EPE-27 Antioxidant Activity Analysis for Polar Extract of Dillena Indica'S Kernel
by: Tania Surya Utami, rit Arbianti, Lina FR
EPE-28
Understanding The Metaspaces of Street Vendors in The Cities: Temporality, Strategies and
Tactics
by: Yandi Andri, Paramita A.
EPE-29 Bringing Sustainability Into The Curriculum of Architectural design Studio
by: Paramita A and Yandi Andri.
EPE-30 LED Technology in Architectural Ligthing of Building'S Faade
by: Siti Hanjarinto, Gregorius A.G. A.
EPE-31 Form in The new Architecture Perspective
by: Yandi Andri Yatmo and Kristanti Dewi Paramita
EPE-32
Utilization Effectiveness of Jakarta Municipal Asset (Case Study: East Jakarta Municipal Youth
Center)
by: Azrar Hadi and Ii Karunia
EPE-33
Water Quality Index Indicator of Water Pullutin. A Case Study in west Tarum Channel as Sources
of Water Supply
by: Djoko M. Hartono, Sulistyoweni W., Dwita Sutjiningsih M.
EPE-34
Non Linear Dynamic Identification of Waste Heat Boiler Unit Based on Adaptive Neuro Fuzzy
Inference System (ANFIS)
by: Yuliati
EPE-35 Load Pattern and Realibility Assessment of Power Generation System
by: Suparman, Zuhal, and Rinaldy D.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-01 1/4
The Effect of Harmonics Current to Performance of
The Over Current Relay
Syafrudin
(1)(2)
and Dahaman Ishak
(1)
(1)
School of Electric and Electronic Engineering Universiti Sains Malaysia
(2)
Fac. of Engineering Electric Department, University of Tanjungpura, Jl. A. Yani, Pontianak
email : syaf_ee@yahoo.com , Syaf@eng.usm.my and dahaman@eng.usm.my
Abstract- The over current relay protection on the
power system is a device that can measure and
responds to abnormal condition and control circuit
breaker to isolate the faulty section of the system.
The over current relays are designed to operate
with sinusoidal current waveform. Now, the
current waveform on the power system have been
distorted non-sinusoidal and lot of containing
harmonic current component, caused by non-liner
loads. For harmonic current condition of system
the performances of that over current relay to be
abnormal.
In this paper, the effect of current harmonic
distortion percentage to over current relay
protection performances is discussed. In this
experiment used an inverse time over current relay
type of induction disc relay. The non-sinusoidal
load currents that consist of different harmonic
spectra were applied to inverse time over current
relay and these nonlinear-load currents were
recorded by a power analyzer instrument
FLUKE 42. According to the experiment results,
the pick up current and the operating time of the
inverse time over current relay increase
proportionally to the total harmonic distortion
(THD) value of the non-sinusoidal current. It is
concluded that, this type of relay cannot protect
the system reliably due to the harmonic
components of current.
Keywords- Non-linear load, Harmonic current,
inverse time over current relay.
I. INTRODUCTION
istorted current waveform content harmonics current
are introduced into power distribution system due to
non-linear loads or system components presenting
non-linear behavior. The increasing use highly non-
linear electric equipment (non-linear loads), the
advancement of semiconductor technology and the
use of more power electronic equipment such as
converters for adjustable speed drives (ASD), switch
mode power supplies, power semiconductor
controller, computers, photocopy machines, TV set
and etc have caused an increasing concern over the
generation of harmonic current and its effects on the
power system. Harmful effect of the harmonic current
on power distribution system like increasing losses,
interferences to communication system, induce
abnormal noise, overheat and burning in various
devices[1].
Harmonic current on power distribution system
also can cause some problems in over current
protection system that provided complete protection
and reliability for the power system. Most of the relay
manufacturers design their relays for pure sinusoidal
current waveform and voltages[2]. The well known
characteristics of relays are not valid under distorted
waveform. Each harmonic frequency component
could produce an independent and cumulative effect,
causing the pick up value of the relay to change
depending on the magnitudes of harmonic
components. Therefore, the relays can not protect the
line or transformer reliably when harmonics are
involved.
In this paper is purposed to determining of the
inverse time over current relay performances for non-
linear load current that containing of different total
harmonic current distortion. In this experiment, non-
sinusoidal wave current that containing harmonic
were applied to the inverse time over current relay and
these currents, pick up current and the operating time
were recorded by a power analyzer instrument.
II. PERFORMANCE OF THE INVERSE TIME
OVER CURRENT RELAY
Inverse time over current relay is used for over
current and short circuit protection of line or
transformer in power distribution systems. In the
Fig.1 shown the construction an electromagnetic
inverse time over current relay type induction disc
unit. All of the operating energy for the relay is
applied to the primary pole coil. The current flowing
in the primary coil will produce a primary flux
U
.
The primary flux will induce an emf in the secondary
coil. The emf in the secondary coil will cause a
current to flow through the winding of the lower
magnet. The secondary current lags behind the
secondary emf. This current creates a magnetic field
in the lower magnet
L
. Both
L
anf
U
will act on
the induction disc and cause it to rotate.
D
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-01 2/4
Fig 1. (a) construction of induction over current relay, (b)
induction disc and contacts (c) moving contact position.
The torque created by the magnetic field is
counteracted by the tension of a spiral spring. When
the turning torque overcomes the force of the spring,
the relay will operate. It is determines the minimum
operating current of relay.
The standard current-time characteristic curve of
inverse time over current relay is shown in Fig.2.
Where the extremely inverse relay saturates at four
multiples of pickup, very inverse at two multiples and
the moderatetly inverse at pickup (IEEE standard
inverse time, 1996:10)[3].
Fig.2 Time-Current curve of inverse time over current relay
The non-sinusoidal current or distorted current consist
of any numbers harmonic current component. Each
frequency is order multiple of the fundamental
frequency. If this current flow into the relays, than
each current harmonic component contributing
increasing frequency. Increasing the frequency of the
input current results in little change in the current that
is produce in the lag coil circuit. However, the flux in
this coil will decrease in inverse proportion to the
frequency increase. Similarly, the flux in the other
coil decreases because of the lowered magnetomotive
force across it. BY decreasing the magnetizing current
for frequency increment and constant lag coil circuit,
the effect is the main coil and the secondary coil
fluxes draw closer in phase. This slow down the disc
rotation, cause the pick up to increase and ultimately
causes the efficiency of the electromagnet to
deteriorate to the point of non-operation. Harmonics
current have serious effect on inverse time over
current relay pic up current value and operation time.
III. EXPERIMENTAL RESULTS
The experiment system configuration as illustrated
in Fig.3.
Fig.3 Experiment system configuration
Current of non-linear loads was isolated by a current
transformer (CT) that has a transformation rate 5A to
5A. All the harmonic amplitude levels and the total
harmonic distortion (THD) recorded by Fluke power
analyzer 42.
The operating time of the inverse time over current
relay is inversely proportional to the current.
Technical information of this relay is given in Table I
below:
Table I. Technical Information of Relay
Frequency 50 Hz
Accuracy of pick up current 5%
Resetting current as % of setting 92%
Contact stop time (after de-
energizing)
0.08s
Consumption at rated current 5VA
Over current adjustment ranges 1.00A-1.25A-1.50A-
2.00A-2.50A-3.00A-
4.00A
As non-linear load for generate harmonic currents is
used a single phase full wave rectifier included with
large variable capacitor C, see Fig 3. For generate
various values of total harmonic current distortion can
do with various values of capacitor C, and with varies
resistive load of rectifiers can be controlled magnitude
of rectifier nput current distorted or current flow
through relay.
The waveform of the non-linear load that contain
harmonic current is shown in Fig 4. This waveform
can be expressed in mathematic form as :
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-01 3/4
=
+ =
,.... 7 , 5 , 3
1 1
) ( sin 2 ) ( sin 2 ) (
h
h h S
t h I t I t i
(1)
where:
i
S
(t)
is distorted current non-sinusoidal
I
1
is current fundamental component
I
h
is current harmonic component orde h
h is harmonic orde ( h = 3, 5, 7, ..)
is angular frequency ( f 2 = )
f is line system frequency
1
is phase angle of current fundamental component
h
is phase angle of current harmonic component orde h
In the rms value form, the current in (1) can be
expressed as:
=
+ =
,.... 7 , 5 , 3
1
h
h S
I I I
(2)
And total harmonic distortion of current is :
1
~
2
2
I
I
THD
h
h
i
=
=
(3)
In this experiment, the rms value and THD of the
current (non-linear loads) are adjusted by changing
the value of capacitor C and resistor of rectifier. The
experiment result as effect of harmonic current to
inverse time over current relay is given in Table II. In
each mode is set for varies of the THD current values
for a constant value pick up current setting of relay.
The true rms current value I
s
, total harmonic distortion
value (THD
i
), the rms value fundamental current I
1
and operating time of relay were measured for each
mode. For all modes, the pick up current of the relay
was adjustable to 1 Ampere.
Table II. Pick up current values and THD
i
values of
relay
Mode THD
i
(%) I
1
(A) I
S
(A)
0 6.00 1.07 1.12
1 32.32 1.14 1.26
2 48.21 1.22 1.34
3 68.99 1.32 1.69
4 71.05 1.38 1.70
5 86.21 1.41 1.93
Table II. Operating time of relay for current is same 2 A
Mode THD
i
(%) I
S
(A) I
S
(A)
0 6.00 2.00 4.71
1 32.32 2.00 6.32
2 48.21 2.00 6.99
3 68.99 2.00 8.82
4 71.05 2.00 9.76
5 86.21 2.00 14.96
According to the experiment results, the pick up
current of inverse time over current relay was affected
by the presence of harmonic current (THD) and the
pick up value was increase by non-sinusoidal currents.
The increment depends on the total harmonic
distortion (THD) value of the non-linear load current.
With respect to the experimental results in Table III,
the pick up current increased as long as THD value of
current was increase. When THDi was above 86%,
the relay which was set to 1 A, operated at 1.9 A.
This shows that the relay can not perform a proper
protection function and causes damage or heating-up
depending on the rms value of the current and process
time in power system components such as
transmission lines, motors and transformers.
The has different operating time value for the
same relay current value in different modes as shown
in Table III. Since the relay induction disc rotates
rather slowly because of non-linear load current
harmonic components, rel;ay operating time will
increase as long as THD value of current increase.
Power system components that are protected by this
relay will be damaged because of the increase in the
operating time.
IV. CONCLUSIONS
Base on experiment results, the inverse time over
current relay induction disc type can not operate
effectively with non-sinusoidal currents that consist of
several harmonic components (THD). Relay pick up
current will increase as long as THD value of the
current increase as shown in Table II. As given in
Table III, the THD of current changes in spite of rms
value of relay. Non-sinusoidal current remains
constant and relay operating time increase as long as
THD value of current increase. Distortion increment
leads to increase in operating time of relay as shown
in Table III. Thus, power system components can
exposed to heat and finally can damaged.
Acknowledgment
The authors wish to thank University Sains Malaysia
School of Electric and Electronic Engineering
MOSTI that have providing the research grant (FRGS
research Grant No 203/PELEC/6071144) and others
facilities which have utilized in this research. The
authors would like to thanks all power group
members-LPKEE-USM.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-01 4/4
REFERENCES
[1] Bishop M.T and Baranowski, J.F, Evaluating
harmonic-Induced Transformer Heating , IEEE
Transactions on Power Delivery, Vol. 10, No.1 305-
310. April 1996.
[2] Elham B.Makram and Regan B Hainnes, Effect of
Harmonic Distortion in Reactive Power Measurement,
IEEE Transaction Ind. Appl, Vol.28, No.4, July/August
1992.
[3] Power System Relaying Committee of IEEE, PES, IEEE
Std.C37112,IEEE Standard Inverse-time
Characteristic equation for over current relays, 1996
[4] Elmore. W.A. Kramer, CA and Zohholl. S.E, Effect of
waveform Distortion on Protective Relays, IEEE
Trans. On Inds. App, Vol 29, No.2, March/April, pp
404-411.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-02 1/5
Effect of Harmonic Current Component to Active Power
Losses on Power Transformer
Syafrudin
1,2
and Dahaman
1
1.
School of Electric and Electronic Engineering Universiti Sains Malaysia
2
Electric Department of Engineering faculty University of Tanjungpura
email : syaf@eng.usm/my
Abstract Power Transformer on the electric power
distribution system used as changer of voltage levels and
also as connector between of the high voltage lines with
the low voltage lines or the consumer voltage
220V/380V. in the recent year, most of loads that
connected on the power distribution system have non-
linear characteristic (non-linear loads) and containing
high current harmonic component, In this paper, the
effect of current harmonic to active power losses on
power transformer is presented. In this case, assumed
that the transformer voltage is sinusoidal without
containing harmonic component and wave form current
transformer is distorted non-sinusoidal and containing
harmonic that generated by non-linear loads. In the
experimental, the transformer active power losses
determined from different value of input and output
power of transformer for linear load with sinusoidal
current waveform without harmonic and also for non-
linear loads with current waveform distorted with
harmonic content THD. This experiment results show
that the transformer active power losses will be increase
proportionally to percentage of the total harmonic
current distortion THD.
Keywords Transformer power losses, non-linear loads,
total harmonic current distortion THD,
I. INTRODUCTION
n the electric power distribution system, power
transformer functioned as a voltage changer from the
high voltage lines or middle voltage lines 20KV-11KV to
low voltage lines or consumer voltage 380V/220V, and also
functioned as loads connected. Generally, loads on the
modern electric power distribution system have the non-
linear current characteristic. It is caused by non-linear loads
such as electrics equipments in the static converter to
electric motor speed drive, computer, system illumination
and lighting use electronic ballast, AC and DC controlled
power supplies, battery charger, etc.
Those non-linier loads cause line current waveform
distorted non-sinusoidal and containing current harmonics
component. Several of the research result expressed that
more than 60% of the loads are connected on the power
distribution system is non-linear loads with total harmonic
current distortion THD content more than 76%, estimated
that this presented will be increase year to years [ 1].
Thereby the loads current that flow trough on power
transformer also distorted non-sinusoidal and containing
current harmonic component. The distortion current
containing harmonic component have high effective value
compared to the current without harmonic sinusoidal [ 2].
Considering the active power losses of the power
transformer consist of the copper power losses in primary
and secondary winding are function of the loads current and
core losses that function of frequency consist eddy current
lose and hysterisis lose, they losses is represented in
thermal form [ 3]. Therefore, the transformer active power
losse depend on total harmonic distortion of loads current
that flow on this transformer.
In this paper, A conducted an experiment to determining of
the effect current harmonic component to active power
losses at one power transformer. Transformer active power
losses are very importance to be known because it is used
for determining of a transformer operation capacities.
Implementation of the transformer active power losses is in
heat form, so that the high power lose can be causing the
transformer over heat and dammed.
Transformer active power losses determined with get
different values between active power input and output that
transformer. This experiment do for transformer with linear
load and also non-linear load. To show the percentage of
total current harmonic effect to active power losses of this
power transformer, so experiment do for any non-linear
load with same active power and varies current harmonic
content.
II. MODELING AND ANALYSIS
A. Harmonic Current Distortion
Generally, ac current waveform which distorted due to
non-linier load is in non-sinusoidal form and periodic wave[
4]. Base on Fourier series analysis, the non-sinusoidal
current waveform will be consisted of one current
fundamental component and any harmonic current
O
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-02 2/5
component that have multiple fundamental frequency. In
mathematic form can be expressed as :
=
+ =
,.... 7 , 5 , 3
1 1
) ( sin 2 ) ( sin 2 ) (
h
h h S
t h I t I t i
(1)
where:
i
S
(t)
is distorted current non-sinusoidal
I
1
is current fundamental component
I
h
is current harmonic component orde h
h is harmonic orde ( h = 3, 5, 7, ..)
is angular frequency ( f 2 = )
f is line system frequency
1
is phase angle of current fundamental component
h
is phase angle of current harmonic component orde h
In the rms value form, the current in (1) can be expressed
as:
=
+ =
,.... 7 , 5 , 3
1
h
h S
I I I
(2)
Percentage of current harmonic content or total harmonic
current distortion THD is defined as bellow :
1
... 9 , 7 , 5 , 3
2
I
I
THD
h
h
h
i
=
= (3)
B. Modeling of Transformer Equivalent Circuit
A power Transformer consist of primary winding,
secondary winding and steel core. Power transformer
equivalent circuit can be draw in figure Fig.1 bellow:
Fig 1. Circuit equivalent of a power transformer
where:
V
p
is primary voltage or input voltage
V
s
is secondary voltage or output voltage
I
p
is primary current or input current
I
s
is secondary current or output current (load current)
R
p
is primary winding resistance
R
s
is secondary winding resistance
X
p
is primary winding reactance
X
s
is secondary winding reactance
X
m
is magnetizing reactance
R
C
adalah resistansi rugi-rugi inti
If the load of transformer is non-linear loads, than the load
current distorted non-sinusoidal and content harmonic
current component expressed as:
=
+ =
,.... 7 , 5 , 3
1
h
hs s s
I I I
(4)
And active power losses of transformer on secondary
winding is :
s
h
h
hs s s s s
R I I R I P
2
... 9 , 7 , 5 , 3
1
2
+ = =
=
(5)
Primary current I
p
consist of secondary current and no-load
current of transformer I
o
, is :
o
h
hs s P
I I I I +
+ =
= ,.... 7 , 5 , 3
1
(6)
While the no-load active power losses consist of or active
power core lose, active power by eddy current and hysterisis
and magnetizing lose. All that losses depend on material of
core and frequency of system. Where the active power lose
of hysterisis is:
=
=
h
h
h hy
f B K P
,... 7 , 5 , 3 , 1
6 . 1
max 1
(7)
and active power lose of the eddy current is :
=
=
,... 7 , 5 , 3 , 1
2
2
max
h
h ed
f B K P (8)
where : K
1
and K
2
is material constants of transformer
Hence, the active power losses on primary side can be
expressed as :
p
h
h
hp p p p p
R I I R I P
2
... 9 , 7 , 5 , 3
1
2
+ = =
=
(9)
Thereby, from equation (5),(7),(8) and (9) hence the
total active power losses a power transformer can be
expressed as bellow:
ed hy S p
P P P P P + + + = (10)
This equation showing that the current harmonic
component is very strong influence to active power losses
value of a power transformer. In this paper, that active
power losses due to harmonic current component of the
power transformer is determined.
III. EXPERIMENTAL RESULTS AND DISCUSSION
A. Experiment
Experiment circuit for determining of active power
losses due to harmonic current on a power transformer is
shown in figure Fig 2. In this experiment do tested to single-
phase power transformer 2000VA, 220V/220V, 50 Hz with
it parameters are bellow :
R
p
= 6.8 , R
s
= 6.8 , L
p
= 0.56 mH , L
s
= 0.56 , X
m
=
1.87 mH.
In this experiment, single phase-transformer is supplied
by single-phase voltage source 120 V, 50 Hz. The fist step,
transformer is given pure resistive or zero THD current with
V
p
R
p
I
p X
p
R
C X
m
N
1
N
2
X
s
R
s
I
s
V
s
I
o
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-02 3/5
a certain power load P
s
. In order the load current flow
trough transformer is sinusoidal without harmonic content.
Then, the power input P
p
and output P
s
of that transformer
are recorded. Active power losses can get from different
values of that power input and power output. The next step,
load changed by non-liner load, it is a single phase full
wave rectifier with smoothing capacitors such as in figure
Fig.3. With the same active power of load, recorded of input
and output THD current and power P
p
and P
s
. Active power
losses of the transformer get from different value of P
p
and
P
s
. This experiment repeated again for several THD current
values. Overall of the experiment results in measurement
data are given in Table I, II, III and IV. For example, in
Figure Fig.4 and 5 also given output currents waveform or
load current with their current harmonic spectrum of this
power transformer. In the Fig.6 show a graph of THD
current content vs transformer active power losses.
Fig 2. Experiment circuit
L N
C
(a)
0
0.005 0.01 0.015 0.025 0.030 0.035 0.040 0.020
V
I
(b)
Gambar 3. (a) Penyearah satu fasa gelombang penuh sebagai
beban non-linier (b) bentuk gelombang arus dan tegangannya
input
Table I. Measurement Data for Linear loads
(Current THD=0%)
Input Output
V
p
I
p
P
p
THD
ip
V
s
I
s
P
s
THD
is
120
V
15.09
A
1.805
KW
0.40 % 116V 10.99
A
1.671
KW
0.40 %
Table II. Measurement Data for Linear loads
(Current THD=32.4%)
Input Output
V
p
I
p
P
p
THD
ip
V
s
I
s
P
s
THD
is
120
V
17.65
A
1.838
KW
31.2% 109V 16.87
A
1.671
KW
32.4 %
Table III. Measurement Data for Linear loads
(Current THD=42.1%)
Input Output
V
p
I
p
P
p
THD
ip
V
s
I
s
P
s
THD
is
120
V
20.31
A
1.853
KW
40.0 % 92V 13.25
A
167
KW
42.1 %
Table IV. Measurement Data for Linear loads
(Current THD=68.0%)
Input Output
V
p
I
p
P
p
THD
ip
V
s
I
s
P
s
THD
is
120
V
24.56
A
1.886
KW
64.30% 86V 15.82
A
167
KW
68.0%
Table V. Measurement Data for Linear loads
(Current THD=81.0%)
Input Output
V
p
I
p
P
p
THD
ip
V
s
I
s
P
s
THD
is
120
V
28.39
A
1.907
KW
76.4% 79V 16.71
A
1.67
KW
81.0%
(a)
Power Analyzer 1
Power Analyzer 2
Trafo
Beban
Non-Linier
Sumber
tegangan
Is
Vs
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-02 4/5
(b)
Fig 4. Output current waveform without harmonic THD=0% and
their spectrum harmonic.
(a)
(b)
Fig 5. Output current waveform without harmonic THD=0% and
their spectrum harmonic.
THD vs Active Power Losses
0
0.05
0.1
0.15
0.2
0.25
0 20 40 60 80 100
%THD Current
A
c
t
i
v
e
P
o
w
e
r
L
o
s
s
e
s
[
k
W
]
Fig 6. Graphic of THD load current vs active power losses
transformer
B. Discussion
Total active power losses of power transformer can be
found from different values of active power input Pp and
output Ps of that transformer. From the Table I,
measurement data to linear load where current flow through
transformer without harmonic {THD=0%), power input of
transformer is P
p
= 1.320 kW and power output is P
s
=
1,099 kW. Hence, active power losses of transformer can
calculated as P = P
p -
P
s
= 0.221 kW. In the same way,
base on data in Table II the active power losses for THD
current of 32.4% is P = 0.134 kW and then from data in
Table III, IV and V for THD current 42.1%, 68.0% and
81.0% respectively, the active power losses of transformer
is with 0.167 kW, 0.182 kW, 0.215 kW and 0.236 kW
respectively.
Normally, power losses a transformer not more than 10% of
total transformer rating. In this case, transformer active
power losses have achieving more than that. Therefore,
increasing THD current of load (non-linear load) can cause
the transformer over heating in the power distribution
system even though the transformer still under rating
operation.
IV. CONCLUSIONS
In this paper have been conducted investigate to a power
transformer to determining of harmonic current effect to
increase the active power losses.
The transformer active power losses can be determine with
different values of input and output active power.
In this experiment, the active power losses increase of the
power transformer is proportional to total harmonic current
distortion THD of transformer loads.
Finally, in order to operate the power transformer for
several operating under THD percentage of non-linear
loads, it is has proper design in terms of harmonic standard,
proper design in terms of loading capacity heating and
insulation lifetime.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-02 5/5
Acknowledgment
The authors wish to thank University Sains Malaysia
School of Electric and Electronic Engineering MOSTI
that have providing the research grant (FRGS research
Grant No 203/PELEC/6071144) and others facilities which
have utilized in this research. The authors would like to
thank all power group members-LPKEE-USM.
REFERENCES
[1] M.Salih TACI and M.Hadi Sarul, The effect of The Harmonic
Components Upon Transformer Active Losses In case of Non-
Sinusoidal Source, IEEE, New York, 2000.
[2] Taci,M.S , Determining The harmonic Effects of Non-Linear
Loads on Parallel Connected Transfoemer in Terms Of Power
Factor, IEEE Power Quality98, Hydrabad-India.
[3] Bishop M.T and Baranowski, J.F, Evaluating harmonic-
Induced Transformer Heating , IEEE Transactions on Power
Delivery, Vol. 10, No.1 305-310. April 1996.
[4] Elham B.Makram and Regan B Hainnes, Effect of Harmonic
Distortion in Reactive Power Measurement , IEEE
Transaction Ind. Appl, Vol.28, No.4, July/August 1992.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-03 1/5
OPTIMIZE PLACING OF PASSIVE FILTER AND
REDUCE HARMONIC IN POWER SYSTEM
Case Study at PT. ISPAT INDO
Iwa Garniwa MK, Budi Sudiarto; Adrianto
Electrical Engineering Department, University of Indonesia
Depok 16424, Indonesia
Abstract-One of the solutions to reduce harmonic
distortion is using passive filter. The most common
type of passive filter is the single-tuned notch filter.
This is the most economical type and is frequently
sufficient for the application. This paper contents
optimizing of placing a passive filter in a power
system so it can be obtain the maximum and most
efficient filterazation to reduce losses in medium
voltage (MV). How to reduce harmonic and improve
power factor using single tuned notch filter is also
studied in this paper. The object of this research is an
power system which has 3 panel of low voltage (LV)
and the each of them has a different non linear loads.
First panel, TR 20, has a maximum THDi (Total
Harmonic Distortion) equal to 7.28 % and maximum
THDv equal to 2.46 % (both in phase 3). Second
panel, TR 21, has a maximum THDi equal to 62.41
% (phase 1) and has maximum THDv equal to 9.5 %
(phase 3). Third panel, TR 23, has a maximum THDi
equal to 58.37 % (phase 1) and maximum THDv
equal to 8.26 % (phase 3). The MV from the
simulation has THDi equal to 34.14 % and THDv
equal to 1.57 %.
I. Introduction
Current from the harmonic load causes extra
heat, isolation failure, operation failure, etc. One of
the solutions to solve this problem is using passive
filter at the source of the harmonic load. Using this
passive filter, harmonic distortion expected to
decrease until the limit of tolerance so that the power
system can work properly.
Research methodology that is been used in
this paper starts with literature study of design
passive filter so it can be implemented to non linear
load. The result of the design filter will be simulated
using ETAP Power Station 4.0.0. Output of the
simulation will be analyzed and then we can make
the best advice to the company which the data is
being used. The company as the object research is
PT.Ispat Indo. PT. Ispat Indo is the biggest steel
company in East Java with the output production
billet and wire rod as the final product. The data were
taken using power and harmonic analyzer equipment.
II. Harmonic and Filter
Harmonic distortion is caused by nonlinear
devices in the power system. A nonlinear devices is
one in which the current is not proportional to the
applied voltage.
Fig 1. Distortion current cause by non linear loads
The sum of sinusoidal waves which create
non sinusoidal wave can be solving using Fourier
analyses.
) 2 sin( 2 ) (
1
0 n
n
n
n
ft n Y Y t Y + =
=
=
(1)
Where
Y
o
= Amplitude of DC current component
Y
n
= RMS value from harmonic component
F = fundamental frequency
n= phase angle from harmonic component
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-03 2/5
Fig 2. Fourier series representation of a distorted
waveform
Total Harmonic Distortion (THD)
expressing distortion level generated by all harmonic
components and defined as follow:
1
1
2
max
M
M
THD
h
n
n
>
= (2)
Where:
THD = Total Harmonic Distortion
M
n
= RMS value of current or voltage n orde
M
1
= RMS value of current or voltage at
fundamental frequency
There are two kinds of filter, active filter
and passive filter. Passive filter is common use in
industry. The weakness using passive filter are
depends on impedance, system frequency, and
component tolerance. Two purposes using filter are:
1. To reduce harmonic in AC network.
2. To provide reactive power for the harmonic
load or other loads.
There are two kind of passive filter, series
filter and shunt filter. Shunt filters that is commonly
use are single tuned filter and order two damped
filter. These two filters are the simplest way of design
and the cheapest to implementation.
Passive filter contents inductance,
capacitance which combines to control harmonic.
Filter passive are commonly use and not expensive.
Using passive filter is possible to interact with the
power system which can disturb it. It is important to
check all of the possibility interaction when filter is
being designed.
The shunt filter works by short-circuiting
harmonic currents as close to the source of the
distortion. This keeps the currents out of the power
system. This filters connected to the bus bar where
reduce of harmonic voltage determined and create a
filter bank.
Fig 3. Passive filter in a power system
There are two parameters need to be
considerate when determine value of R, L, and C:
a. Quality factor (Q)
b. Deviation relative frequency ()
III. Measurement and Data
Program we use to simulate the power
system is ETAP 4.0.0.
Fig 4. Power system design using
ETAP 4.0.0
The data has been recorded using 9625
Power Measurement and transferred to the personal
computer using 9625 Power Measurement Support
Software. The data of MV we have it from the
simulation software (ETAP 4.0.0).
Table 1. Data from the panels when
the THDi maximum
TR 20 TR 21 TR 23 TM
THDi 7.28% 62.41% 58.37% 34.14%
Current 1344 313.3 A 372.5 A 92 A
THDv 2.37% 3.70% 3.42% 1.57%
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-03 3/5
Power
Factor
0.68 0.68 0.7 0.66
(a)
(b)
Fig 4. (a) Current distortion wave in TR 20, 21, and
23
(b) Current distortion spectrum in TR 20, 21, and 23
Using IEC 61000-3-6 and 61000-3-4
standard, the limit of THDv is 8 %. From the data,
we can determine the panels which need a passive
filter are TR 21, TR 23, and MV. These panels have
THDv and distortion current over the limit.
IV. Analyze and Solution
By using the equation below, we can
determine the losses in power system:
R I P
losses
2
= (3)
Where
I = total current
R = Impedance
P = power loss
If the impedance of the system constant, we can
decrease the losses by decreasing the current.
Single tuned notch filter is use to reduce
harmonic. The basis parameters of design filter as
follow:
Table 3. TR 21 Filter harmonic data
THDi Maximum (%) 76 Phasa 1
THDv with highest THDi
(%) 4.46
THDv with highest
harmonic order (%) 3.3 Order 5
PFCC kVAR ON 166
Table 4. TR 23 Filter harmonic data
THDi Maximum (%) 68 Phasa 1
THDv with highest THDi
(%) 4.75
THDv with highest
harmonic order(%) 3.56 Order 5
PFCC kVAR ON 189
Table 5. TM Filter harmonic data
THDi Maximum (%) 34 Phasa 1
THDv with highestTHDi
(%) 1.47
THDv with highest
harmonic order (%) 1.23 Order 5
PFCC kVAR ON 878
The steps to design a filter based on the
table above as follow:
1. Election of frequency filter.
From the measurement, the highest
harmonic level happened at order 5. Tuning of
order degraded by a little below order in fact so
that n = 4.8. This degradation is conducted as
tolerance to filter component and also variation
of system impedance.
2. Enumeration of capacitor size measure and
resonance frequency.
The formula of current at the fundamental
frequency for capacitor bank as follow:
actual
actual
FLcap
kV
kVAR
I
3
= (4)
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-03 4/5
Formula to look for capacitive impedance as
follow:
rated
rated
C
M
kV
X
var
2
= (5)
For TR 20, nominal voltage phase to phase is
430 V. For TR 21 and 23 is 730 V and for MV
is 11 kV.
Inductor impedance determined with
formula following:
2
n
X
X
C
R
= (6)
Because there is reactor, the fundamental
current increase which can expressed by:
( )
R C
bus
FLfilter
X X
V
I
=
3
(7)
Because the current increased, the reactive
power compensation increase and we can
determine using:
FLfilter bus plied
I V kVAR * * 3
sup
= (8)
Table 6. Filter specification
THDi Maximum Filter
Spesification
TR 21 TR 23 TM
System
frequency 50 50 50
Rating of
Capasitor Bank 166 189 878
Rating of Bank
Capacitor
Current 131.25 150.45 46.11
Tuning
harmonic filter 4.8 4.8 4.8
Tuning system
frequency 240 240 240
Capacitor
Impedance 3.211 2.801 137.742
Reactor
Impedansi 0.139 0.122 5.978
Filter Full Load
Current 137.19 157.3 48.2
There are 4 variations when placing the
filter. The results of simulation using ETAP 4.0.0 in
TM as follow:
Table 7. Filter at TR 21
Before After
Difference
THDi (%) 34.14 28 6.14
THDv (%) 1.57 1.26 0.31
Current
(A) 92 84 8
Power
factor 0.66 0.71 -0.05
Table 8. Filter at TR 23
Before After
Differense
THDi (%) 34.14 27 7.14
THDv (%) 1.57 1.21 0.36
Current
(A) 92 83 9
Power
factor 0.66 0.72 -0.06
Table 9. Filter at TR 21 and TR 23
Before After
Difference
THDi (%) 34.14 22 12.14
THDv (%) 1.57 0.98 0.59
Current
(A) 92 76 16
Faktor
Daya 0.66 0.77 -0.11
Table 10. Filter at TM
Before After
Difference
THDi (%) 34.14 27 7.14
THDv (%) 1.57 0.98 0.59
Current
(A) 92 62 30
Power
factor 0.66 0.97 -0.31
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-03 5/5
From the result of simulation above, we can
determine that the biggest difference in decreasing
current is when the filter located at TM. The current
was decreased 30 A to the final value 62 A. This
happen because harmonic distortion was reduced by
the passive filter which the filter designed based on
the entire load (TR 20, TR 21, and TR 23). This
makes the filter design more precision. Reactive
power that is supplied by the capacitor bank also
based on the entire load so its enough to increase the
power factor value to 0.97 which can avoid the
penalty. This also helps make the losses in TM
decrease. The THDv was decreased to 0.98 % from
the original value 1.57 %. The THDi also decreased
to 27 % from the original value 31.14 %. These data
shows us that passive filter can reduce harmonic,
increase power factor, and reduce the power loss.
V. Conclusion
1. Single tuned filter can be use to reduce
harmonic distortion, increase power factor, and
reduce power loss.
2. Optimize placing of passive filter can make
passive filter more efficient and effective.
3. Placing the passive filter at the MV in the object
of this research is the best place to minimize the
power loss.
4. The results using filter at MV are:
THDi = 27 %
THDv = 0.98 %
Current = 62 A
Power factor = 0.97
VI. Reference
1. Roger C. Dugan..[et al]. 2002. Electric Power
System Quality. New York : McGraw-Hill
2. Cahyadi, Rifky. 2003. Upaya Menghemat
Energi dengan Mereduksi Harmonisa. Depok :
Skripsi Departemen Teknik Elektro Universitas
Indonesia
3. Setiabudy, Rudy. 2006. The Design Of Passive
Filter To Overcome Harmonic Distortion.
Depok : Proceedings. IJJSS 2006
4. Zunaedi, Totok. 1999. Eliminasi Harmonik Dan
Kompensasi Daya Reaktif di PT. Engenys Steel.
Depok : Skripsi Departemen Teknik Elektro
Universitas Indonesia.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-04 1/5
CONVERGENCE APPROXIMATION OF ABR FORMULATION
FOR JOSEPHSONS TUNNELING IN UO
2
CHAIN REACTION
AT 45.7 MW CANDU NUCLEAR REACTOR
Moh. Hardiyanto
Department of Educational Physics FTMIPA - UNINDRA
Large Hadron Collider Laboratory, CERN Lyon, France
E-mail : moh_hardiyanto_iti@yahoo.com
Abstract - The convergence of free covariant
equation in Einsteins space with quantum condition
is studied using the ABR (Abrikosov-Balseiro-
Russell) formulation in convergence approximation
for Josephson tunneling is important role for
determine of neutrino particle existing, especially
after Cerenkovs effect for 45.7 megawatts nuclear
reactor based on UO
2
chain reaction. This
approaching will be solved the problem for determine
the value of interstellar Electrical Conductivity (EC)
on UO
2
chain reaction, then the post condition of
muon has been known exactly. In this research shown
the value of EC is 4.32 eV at 378 tesla magnetic field
for 2.1 x 10
4
currie/mm fast thermal neutron floating
in 45.7 megawatts adjusted power of Canadian
Deuterium Uranium (CANDU) nuclear reactor. The
pictures had resulted by special Electron-Scanning-
Nuclear-Absorbtion (ESNA) shown any possibilities
of Josephsons tunneling must be boundary by muon
particles without neutrino particle existing for 350
456 tesla magnetic field on UO
2
more enrichment
nuclear fuel at CANDU, whereas this research has
purpose for provide the mathematical formulation to
boundary of muons moving at nuclear research
reactor to a high degree of accuracy and with Catch-
Nuc, one of nuclear beam equipment has a few
important value of experimental effort.
Keywords - ABR formulation, EC value, UO
2
chain reaction, magnetic field value
I. INTRODUCTION
Its is known that for muon-hadron
scattering, the close-coupling equations [1] have
been used extensively in the sophisticated
computation in Abrikosov-Balseiro-Russell (ABR)
formulation [2]. In the close-coupling equations,
the complete wave functions are expanded in target
states on Einsteins space. In this paper the target
states are constructured in the finite L
2
basis space
following the results of Abellian system and next
formulated by ABR without Diracs condition. The
previous studied the convergence of the
approximation target states for discrete and
continuum cases in quantum condition for UO
2
chain reaction at 45.7 megawatts (MW) adjusted
power in Canadian Deuterium Uranium (CANDU)
nuclear research reactor. The advanced studied its
application to the ABR equations which results in
pseudo state close-coupling approximations for
bound-free transition in convergence structure for
Josephsons tunneling. For first step using by
Einsteins space for eliminary construction of ABR
formulation in UO
2
chain reaction.
In this paper, the Abellian system was
conducted for covariant equation to Einsteins
space and ignores the Diracs condition for 300
410 tesla magnetic fields on fast thermal neutron,
to study the convergence behavior of the free
covariant equation part. Its possible analytical
computation is also addressed after finding the
asymptotic behavior.
II. Pseudostate Close-Coupling Approximations
The close-coupling equations in ABR
formulation require to study is given by
[ ]
=
= +
1
0
k
k jk j j
f V f E H (1)
where
2
0
2
1
= H (2)
the total energy of the system connecting the initial
and final state f i satisfies
2 2
1
2
1
2
1
f f i
k k E + = + = (3)
j
f are channel functions related to the
complete wave function as
) ( ) ( ) , (
2
1
1 2 1
r f r r r
j
j
j
= (4)
in which the + (-) superscript refers to singlet
(triplet) scattering and ) (
1
r
j
are muon eigen
functions.
The channel potential
jk
V is given by
+ =
jk jk jk
W U V (5)
where
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-04 2/5
+
> =
2 1
2 1
1
2
2 1 1
2 1 2 1
) (
1
) (
1 1
) (
r r
r r
r
r
r r r
r r dr dr U
k
k
j jk
(6)
and
i i j j
i
i jk jk
E W ) ( + =
(7)
with
1 =
i
1
4
=
i
(8)
If replace the target states by a set of pseudo states
(as consequences of finite basis set) in order to
solve the equations approximately. We has to
modify the close-coupling equations for Einsteins
space through for normal condition after
Cerenkovs effect at 315 tesla magnetic field since
the exchange term in the form (7) is valid only for
exact target states. The form appropriate to pseudo
states is given by
Nj Nk Nj Nk jk
r H r H W >< + >< =
) ( ) (
2 1
< > +
N
i
Ni i Ni jk
E
1
) ( (9)
where
i N
label the pseudo states and H(r
1
) is the
target non-linier for Hamiltonian operator in
Abellian.
In calculating the pseudo state close-
coupling approximation one must consider the
error present in the quadrature rule approximation
and the close-coupling potentials when pseudo
states are employed. The first error will not be
discussed here. The second error will be discussed
in association with the direct-potential component
of the channel potentials since they are regarded is
being responsible for the dominant scattering
process especially at higher energy.
III. Convergence Approximation of ABR
Formulation
In the momentum re-presentation, the
derivative of ABR formulation for muon state in
UO
2
chain reaction is given by
V G V V T
B 0 2
+ = (10)
Here G
0
is the diagonal matrix of free channel
Greens functions. This second ABR formulation
approximation has been used by Einsteins space in
Abellian system, especially the fast thermal neutron
floating at 350 tesla magnetic field to test the
suitability of pseudo states expansions with initial
and first states chosen either the ground or 2s, 2p
excited states.
The actual Josephson tunneling element
from (10) require to study is
[ ]
+
=
dr r e r
k k V
j
r i
i ij
j i ij
) ( ) (
2 ) , (
2
(11)
where
j i
k k = . The indices i and j can be
either discrete or continuous after Cerenkovs
effect on 2.1 x 10
4
currie/mm and the range of 315
390 tesla magnetic field. For free covariant
equation in quantum condition for UO
2
enrichment
reaction potentials one uses
= dr r e r I
p
r i
q pq
) ( ) ( 2 ) (
. 2
(12)
where
p
and
q
are initial and final continuum
states having the momenta p and q respectively,
given by
=
|
|
\
|
+
=
0
.
) ( ) 1 2 (
2
1
) (
q
N
q
i
q
q
qr
r q
P r U e i
qr
r
q q
q q
l
l l
l l
l
(13)
and normalized to a function in
3
) 2 /( q . The
approximate target wave function
N
q
p
U is given by
q q
q q
n n
q
N
n
N
q
x p
n
n
q B U
l l
l l
l
) (
2 2 (
) 1 (
) (
1
1
0
+
=
+ +
+
>=
(14)
where
) 2 / (
2 / ) 1 (
2
) 1 (
) 1 ( 2 ) (
+ =
e x
i q B
q
q
q
p
q
p
l
l
(15)
and
..... 0 .
) ( ) ( ) (
1 2
2 /
1
=
=
+
+
q
n
r
n
n
r L e r r
q q q
l
l l l
(16)
solving the integral (15) after quite lengthy
derivation one finally obtains
) (
. .
, .
1
2
3
) 1 (
) ( 32 ) (
q
p
q p q
q p
q p
i
L m
m m M
L
pq
c i
pq V
l
l
l l
l l
+ +
) (
0 00
) 1 2 ( 4
) 1 2 )( 1 2 (
2 / 1
+
+ +
L
pq q p q p
q p
q p
I LM m m
L
L
x
l l
l l
l l
(17)
where
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-04 3/5
) ( ) ( ) (
) (
) 2 2 ( ) 2 2 (
) (
1
1 , 1
0 .
2
q m p n
M N
m n
q
q p
L
pq
x P x P q B
p B I
p p
p
q p
l l
l
l l
l
l l
+
=
+ +
+ +
=
)
`
+
+
) ( ) 2 ( Re
1 1
i i i x
v L
) . : 2 2 . 2 2 :
2 ( ) 2 (
2
+ +
+ +
+ + + +
i i
m n v F v x
q p
q p q p
l l
l l l l
(18)
IV. METHODOLOGY
This research using by high degree
accuracy mathematical approaching and a few
experimental efforts with special Electron-
Scanning-Nuclear-Absorbtion (ESNA) and Catch-
Nuc, for see the results of the mathematical
approximation, in case is ABR formulation in
Josephsons tunneling for UO
2
chain reaction at
350 456 tesla magnetic field and determine of
Electrical Conductivity (EC) value in 2.1 x 10
4
fast
thermal neutron floating.
A few mathematical equations in quantum
condition will be explain in this section.
ABR Formulation non Abellian
The metric tensor is defined through the
line element ds as follows:
dx dx g ds =
2
(19)
Hence, different metric shall lead to different
properties of space-time. For a space time where
the metric is defined through
( ) [ ]
|
|
\
|
+ +
=
) sin (
1
2 2 2 2
2
2
2 2 2
d d r
kr
dr
t R dt ds
(20)
the space satisfies the properties that it is
homogeneous and isotropic. Such properties agree
with the nuclear structure in microscopic principles
and are also supported by experimental data.
Accordingly the above metric, called the ABR
Formulation without Abellian system, becomes a
standard model of UO
2
structure. In the above
expression R(t) denote the scale factor and k is a
constant. The universe is closed if k > 0, open if k <
0 and flat if k = 0 and R(t) sometimes are rescaled
in such a way that the value of k assume one of the
three values -1, 0, or +1. The Cartesian form of the
above line element.
+ +
+
+
+
=
2
2
2
2
2
2
2
2
2
2
2 2
1
1
1
1
1
1
) (
kr
kz
dz
kr
ky
dy
kr
kx
dx
t R
dt ds
(21)
|
|
\
|
x
g
x
g
x
g
g
2
1
(22)
Combining (19) and (21), one finds the explicit
forms of non-zero components of the Riemannian
curvature tensors as:
. 00
) (
) (
ab
a
b
t R
t R
R =
+ =
2
0
0
1
) ( ) (
kr
x kx
t R t R R
b a
ab b a
(23)
|
\
|
+ =
A A dA A Tr
k
S
M
3
2
4
(24)
) 2 ( ) 1 (
0 0
' ) , 2 , 1 ( '
E E K
=
(25)
( ) =
3
1
2
24
1
M
wZ
dg g Tr S
(26)
=
1
0
) ( ) ( da ax F ax x A
(27)
=
1
0
) ( ) ( x R d x x
(28)
0
1 '
) (
2
1
) (
'
) 2 (
^
E q g
n
s
Q Q
s
s
l
l
=
|
|
\
|
=
(29)
V. RESULT AND DISCUSSION
It is shown in the final equation that beside the
double series n and m which are finite, the ABR
function F
2
is formally defined by a double series
expansion with radius of convergence severely
restricted. Because the occurrence of n and m
which are finite in the arguments, there is no
problems arise about its convergence. The
remaining is to compute (22) numerically in order
to see how rapid the convergence is since the
expression of Riemannian curvature tensor is
similar with the expression of
F in the non
Abellian gauge. Such a formula (18) may also be
obtained in the general theory of relativity, the fast
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-04 4/5
thermal neutron break into UO
2
matrix is show in
Fig. 1.
Fig.1. The floating of muon moving
in UO
2
matrix at 378 tesla
magnetic field
(Courtesy of CANDU nuclear reactor, Canada)
To obtain this formula, let us impose the ABR
gauge condition on the Christoffel symbol as
follows:
0 ) ( = x x
a
(30)
If we make the transformation of
x x
where : [ ] 1 , 0 is a parameter could obtain:
) (
) ( ) (
x
x x x R x
a
+
=
(31)
and if use the relation:
). (
) (
) (
) ( ) (
) (
x F x
x
x F
x
x
x F
d
x d
x F
d
d
=
(32)
)) ( ( ) ( x
d
d
x R x
= (33)
Using the above formula, we shall derive
the Christoffel symbol and contrast them with the
Christoffel symbol previously derive from the ABR
formulation on Abellian system and the re-
polarization of UO
2
matrix will give in Fig. 2.
Fig.2. The floating of muon moving
in UO
2
matrix at 378 tesla magnetic field and fast
thermal neutron at 2.1 x 10
4
currie/mm floating
(Courtesy of CANDU nuclear reactor, Canada)
We have seen that the ABR gauge which
satisfy the Josephsons tunneling effect can be
deduced from (23). We shall refer the symbol of
FS
+
+
+
+
+
+
+
= + =
+
+
)! 4 (
) 2 ( ) 1 (
)! 3 (
) 1 ( 2 ) 1 (
)! 2 (
) 1 (
) ( 0 ) , ( ) , (
2
1
) (
n
n
n
n
n
r F kn t r B t r E
n
n n
n
(34)
Fig.3. The floating of muon moving
in UO
2
matrix at 378 tesla magnetic field and fast
thermal neutron at 2.1 x 10
4
currie/mm floating without
neutrino particle
(Courtesy of CANDU nuclear reactor, Canada)
Based on Fig. 3, if integrate the equation by parts,
we will obtain the series form of the equation as
=
+
=
1
0
0
)! 2 (
) 1 (
) (
n
n
n n
x
d
d
n
x R x d
(35)
Let us now choose a scale factor as:
R(t) = e
mt
. This is an asymptotic value of the
Friedmann model of UO
2
nuclear structure with
3 / = m , where is the UO
2
constant.
Using by special ESNA, we get the
Josephsons tunneling with Abellian operator such
as the Fig. 4.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-04 5/5
Fig.4. The Josephsons tunneling
in UO
2
matrix at 378 tesla magnetic field
(Courtesy of CANDU nuclear reactor, Canada)
The singularity for the first turns out to be worse
compared to the later since the standard one has
only a singularity factor from EC value at
interstellar of muons moving for 2.1 x 10
4
fast
thermal neutron floating at range 350 456 tesla
magnetic field without neutrino particle existing
after Cerenkovs of ( ) V e kr 32 . 4 1 / 1
2
= .
At least on the pedagogical point of view one then
becomes aware that the ABR formulation in UO
2
chain reaction for a given Riemannian curvature
tensor are not unique.
VI. CONCLUSIONS
Investigations and research using by
ABR formulation in convergence approximation
and ESNA also Catch-Nuc equipments based on
UO
2
chain reaction by Josephson tunneling at 45.7
MW CANDU nuclear research reactor has a few
result, expressed:
a. The strength of fast thermal neutron
floating is 2.1 x 10
4
currie/mm.
b. The values of Electrical Conductivity
(EC) for interstellar muons moving is
4.32 eV at 378 tesla magnetic field.
c. These equations have a high degree
accuracy, so they could be determine of
muons moving after Cerenkovs effect
for 45.7 MW adjusted power.
ACKNOWLEDGMENT
The author wish to thank Canadian
Deuterium Uranium (CANDU) Nuclear Reactor
team for contributions to this work.
REFERENCES
[1] Bransden,B.H.Scott,T.,Shingal,R.
and Roychoudhury,R.K., Journal
Phys.B 15, 2003
[2] Bray I. and Stelbovics.A.T., Phys.Rev. A. 46,
2002
[3] Burke. P.G and Seaton M.J.,
Math. Journal of Phys.10 , 2002
[4] CANDU nuclear reactor, UO
2
Interference
Structure, Canada, 2003
[5] Flinder, and Stelbovics.A.T.,Test of Close-
Coupling in the poet Model, Procedings of the
Sixth International Symposium on Correlations
and polarization in Electronic and Atomic
Collisions and (e,2e) Reactions, Univ. of
California, USA (2001).
[6] Helna, Duprix, Structure of ABR
formulation in Diracs condition, Universit
du Qubec, Canada, 2003
[7] Madison,D.H.and Callaway.J.J.,
Journal Phys B 20, 2002
[8] Stelbovics,A.T., and Abrikosov G.,
American J.Phys.43, 2003
About Author:
Lecturer of Department of Educational Physics
and Researcher at Large Hadron Collider
(LHC) Laboratory Lyon - France, and now
take a doctoral program at Universite du
Quebec, Canada.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-05 1/4
Energy Conservation on Steel Industry
by Increasing Efficiency and Electric Power Quality
Iwa Garniwa; Budi Sudiarto; Fauzan Hanif Jufri
Faculty of Engineering, University of Indonesia
e-mail : iwa@ee.ui.ac.id ; budi@ee.ui.ac.id ; fauzantolenk@yahoo.com
Abstract Electrical energy conservation is
electric energy usage with high efficiency by
minimize losses at all production process, start
from generation, transmission, distribution, and
consumption.
Industry sectors included steel industries are
large energy consumers because industry sectors
use about 30% primary energy. This mean energy
conservation on industry sector gives a significant
role for thrift the energy.
This paper did some research for energy
conservation opportunity on demand side focused
in electrical energy conservation, especially in
compact energy industries, that is steel industries.
This research takes the some opportunity by
increasing efficiency on processes, power systems,
and electric power quality on electrical systems.
Energy consumption efficiency for every process
can be explained from Specific Energy
Consumption (SEC), that is a value meant ratio
between total energy consumption to total product
yield.
The recommendations yield from this paper
will be followed by feasibility study from technical
sides and economic sides, using Life-Cycle Costing
Analysis. The opportunities can be high cost or
investment, or low cost or no cost.
The data on this paper get from survey to one
of steel industries, that is PT SERMANI STEEL,
which produce galvanized iron sheet, and PT
SURABAYA WIRE which produce nails and wire
rod.
Keywords Energy Conservation, steel industry,
electric power quality, energy efficiency
I. INTRODUCTION
ndustry sectors are large energy consumers because
they use about 30% primary energy. This mean
energy conservation on industry sectors give a
significant role for thrift the energy. Besides that,
industry sector consist of large consumers which a
small number, so energy conservation can be easier
than do at other consumers. Energy sectors can be
classified based on its energy consumption, those are
compact energy, for example steel industries, cement
industries, and melting factory.
Processes on steel industries use much energy.
Those are used for electric motor, convert electrical
energy to mechanical energy, and for arc furnace,
convert electrical energy to heat energy. Electrical
energy usage can be reduced by increase or modify
another factors used in processes. As an example,
increasing power factor, efficient loading on motors
and transformers, or increasing the fluid pressure on
furnace, etc.
II. BASIC THEORY [1]
Energy intensity in producing products defined as
Specific Energy Consumption (SEC), it is ratio
between energy consumed to yield products to
amount of products yield. The unit is kWh/Ton. By
conserve energy, we hope that the SEC will decrease,
which mean decrease the energy consumption without
decrease the product yield.
2.1 Improving power factor
The reactive power has to be supplied by the
utility even though it is not actually registered by the
power meter (as real power used). The magnitude of
this reactive power increases as the power factor
decreases. To account for the loss of energy due to the
reactive power, most utilities have established rate
structures that penalize any user that has low power
factor. Therefore, significant savings in the utility
costs can be achieved by improving the power factor.
This improvement can be obtained by adding a set of
capacitors to the entire electrical system. See figure 1,
Figure 1. Improving power factor by using capacitor
The size of these capacitors (Q compensator) is
typically measured in KVAR (the same unit as the
reactive power) and can be determined, as shown in
figure 1, using the power triangle analysis :
I
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-05 2/4
tan tan
1 1
tan cos tan cos
P Q Q P
C ex comp ex comp
P P PF PF
C ex comp
| |
= = |
\
| |
= |
\
(1)
Where P stands for the real power (measured in kW),
while Qe and Qc stand for the reactive power before
and after the retrofit, respectively.
The thrift yield from improving power factor can
be defined as [2] :
2
% 1 100%
ex
reduction
comp
PF
Loss
PF
| |
= |
|
\
(2)
2.2 Reducing harmonic distortion
Harmonic distortion (based on IEC 702-07-43) is
waveform distortion which frequency multiplied from
its fundamental frequency in integer number.
Figure 2. Harmonic distortion waveform
To quantify the level of distortion for voltage and
current a dimensionless number, referred to as the
total harmonic distortion (THD), is determined
through a Fourier series analysis of the voltage and
current waveforms and is respectively defined as:
2 2
1
1
1
h
R M S M M T H D
h
h
= =
>
(3)
Harmonic distortion will increase the RMS value
both current and voltage, so that will increase power
loss in the systems. Besides that, harmonic distortion
will reduce equipment lifetime, three phase neutral
current overload, etc.
2.3 Replacement motor with energy-efficient motor
Induction motor commonly used in industrial
sectors has an inductive characteristic, produced by
its magnetic field to rotate the rotor.
One parameter is typically important to identify an
electric motor during full-load operation, the
conversion efficiency of the motor (). This efficiency
expresses the mechanical power as a fraction of the
real electric power consumed by the motor.
mechanical
electrical
P
P
= (4)
Due to various losses (friction, core losses due to
the alternating of the magnetic field and resistive
losses through the windings), the motor efficiency has
typical values ranging from 75% to 95%, depending
on the size of the motor. In the above definition, PM
stands for the mechanical power output of the motor,
expressed in kW or horsepower (HP), which is the
most important factor in selecting a motor.
Table 1. Typical motor efficiencies (Hoshide, 1994)
The improved efficiency of the high/premium-
motors is mainly due to better design with use of
better materials to reduce losses, which however
comes with a higher price (about 10 to 30% more than
standard-efficiency motors). This fact partially
explains why only one-fifth of the motors sold in the
US are energy-efficient.
2.4 Use Adjustable-Speed Driver (ASD)
The use of adjustable-speed drivers (ASDs) offers
numerous advantages to drive systems' energy
efficiency. They can offer the potential for higher
efficiencies, lower operating costs, easier control and
minimal maintenance. The most frequently used type
of adjustable-speed drivers is the frequency controlled
ASDs offering the greatest advantages. They consist
of two types: direct and indirect inverters. The first
ones convert frequency and voltage in one step,
whereas the latter use an intermediary DC link. ASDs
can take the saving about 50% in reheating furnace
2.5 Reducing peak of power demand
This method used by controlling the demand in
consumer side. For reducing peak of demand power, it
can be determined by demand factor and load factor.
If demand factor is low, it means that the connected
load is not efficient. Demand factor describe as
maximum demand
demand factor =
connected load
(5)
Life-Cycle Cost Analysis [3]
LCC = I + R - R + E + OM&R
cpl es
(6)
[ ] NS = E + W + OM&R - I + R - R
A:BC 0 cpl es
(7)
E + OM&R
SIR =
A:BC
I + R - R
0 cpl es
(8)
( ) ( )
1
AIRR = 1 + r SIR -1
N
(9)
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-05 3/4
( )
E + OM&R - R + R
t t cpl es
t t
0
1 1
I
t
d t
+ =
y
(10)
Where,
LCC = present value of LCC
I = present value of investment
R
cpl
= present value of capital replacement cost
R
es
=present value of residual cost
E = present value of energy cost
OM&R = present value of operation, maintanance,
repairement cost
:
NS
A BC
= Net Saving
E =
( )
E - E
BC A
= energy cost saving
III. EXPERIMENTAL RESULTS
3.1 PT SERMANI STEEL
The SEC of PT SERMANI STEEL before energy
conservation done is :
KES
Bulan kWh Lembar kG
kWh/Lbr kWh/kG
September 2006 58.600 333.768 1094.322 0,18 0,05
Oktober 2006 63.400 206.265 667.272 0,31 0,10
November 2006 50.900 321.493 1.027.163 0,16 0,05
Desember 2006 61.400 314.480 998.718 0,20 0,06
Januari 2007 60.300 371.483 1.135.617 0,16 0,05
Februari 2007 69.200 315.667 998.460 0,22 0,07
Maret 2007 59.000 310.368 1.053.246 0,19 0,06
April 2007 67.100 188.735 647.882 0,36 0,10
Mei 2007 47.400 354.857 1.135.782 0,13 0,04
Juni 2007 69.900 226.372 722.879 0,31 0,10
Juli 2007 53.600 342.927 766.234 0,16 0,07
Agustus 2007 47.400 119.191 417.582 0,40 0,11
Rata-rata 59.016.67 283.801 888.763,1 0,23 0,066
Table 2. SEC of PT SERMANI STEEL before energy
conservation
The energy conservations which passed by LCC
analysis are :
energy
saving
investment
average
saving
pay back
periode
(kWh/month) (Rp) (Rp/month) (month)
Reducing
connected load
0 13,935,000 2,550,428 3
Improving
power factor
625.68 20,000,000 486,084 20
energy
conservation
Table 3. Electrical energy conservation of PT SERMANI
STEEL
Using ASDs and replacing the motor did not pass
the economic requirement, so they are not feasible for
current time. However, if there will be a plan for
reconstruction and has big enough money for
investment, it will be best to use ASDs and premium-
efficiency motor. While harmonic distortion is just
about 2 % 3 %.
The new SEC after energy conservation are :
before energy
conservation
after energy
conservation
saving
Product yield (Ton) 888.76 888.76
Product Cost (Rp) 54,458,719.00 51,422,207.00 3,036,512.00
Energy Conumption (kWh) 59,016.00 58,390.12 625.88
Rp/Ton 61,274.94 57,858.37 3,416.57
SEC (kWh/Ton) 66.40 65.70 0.70
SEC (kWh/kG) 0.0664 0.0657 0.000704
Table 4. SEC of PT SERMANI STEEL after energy
conservation
From table 4, SEC of PT SERMANI STEEL
decrease 0,7 kWh/ton a month.
3.2 PT SURABAYA WIRE
The SEC of PT SURABAYA WIRE before
energy conservation done is :
Bulan kWh kG
KES
(kWh/kG)
Agustus 2006 160.500 2.187.689 0,07
September 2006 206.100 3.286.985 0,06
Oktober 2006 244.500 2.128.311 0,11
November 2006 173.790 3.703.988 0,05
Desember 2006 284.940 2.908.885 0,10
Januari 2007 167.040 2.417.828 0,07
Februari 2007 190.950 2.158.409 0,09
Maret 2007 173.220 2.941.461 0,06
April 2007 159.690 1.714.668 0,09
Mei 2007 88.620 1.790.362 0,05
Juni2007 139.110 2.593.985 0,05
Juli 2007 181.560 2.216.294 0,08
Rata-rata 180.835 2.504.072 0,07
Table 5. SEC of PT SURABAYA WIRE before energy
conservation
The energy conservations which passed by LCC
analysis are :
energy
saving
investment
average
saving
pay back
periode
(kWh/month) (Rp) (Rp/month) (month)
reducing harmonic
distortion
14,291,280 25,000,000 3,488,358 7
improving power
factor
4,087,459 40,000,000 654,139 24
energy
conservation
Table 6. Electrical energy conservation of PT SURABAYA
WIRE
Demand factor of PT SURABAYA WIRE is
90,1% so that the connected load could not be
reduced. Usage of ASDs and replacement of
premium-efficiency motor do not pass the LCC
Analysis for current time.
The new SEC after energy conservation are :
before energy
conservation
after energy
conservation
saving
Product yield (Ton) 2,504 2,504 -
Product Cost (Rp) 136,364,807 126,256,500.55 10,108,306.45
Energy Conumption (kWh) 180,835 162,456.26 18,378.74
Rp/Ton 54,457.22 50,420.48 4,036.75
SEC (kWh/Ton) 72.22 64.88 7.34
SEC (kWh/kG) 0.0722 0.0649 0.0073
Table 7. SEC of PT SURABAYA WIRE after energy
conservation
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-05 4/4
From table 7, SEC of PT SURABAYA WIRE
decrease 7,34 kWh/ton a month.
IV. CONCLUSIONS
T Th he er re e a ar re e f fi iv ve e m me et th ho od ds s t to o c co on ns se er rv ve e e el le ec ct tr ri ic ca al l
e en ne er rg gy y i in n i in nd du us st tr ri ia al l p po ow we er r s sy ys st te em m, , t th he ey y a ar re e i im mp pr ro ov vi in ng g
p po ow we er r f fa ac ct to or r, , r re ed du uc ci in ng g h ha ar rm mo on ni ic c d di is st to or rt ti io on n, , u us si in ng g
A Ad dj ju us st ta ab bl le e- -S Sp pe ee ed d D Dr ri iv ve er r, , r re ep pl la ac ci in ng g m mo ot to or r w wi it th h h hi ig gh h- -
e ef ff fi ic ci ie en nc cy y o on ne e, , a an nd d c co on nt tr ro ol ll li in ng g p pe ea ak k o of f p po ow we er r
d de em ma an nd d. . T Th he ey y c co on ns si is st t o of f h hi ig gh h c co os st t m me et th ho od d a an nd d l lo ow w
c co os st t m me et th ho od d. . H Hi ig gh h c co os st t m me et th ho od d n ne ee ed d a an n a ad dv va an nc ce e s st tu ud dy y
t to o t ta ak ke e t th he e d de ec ci is si io on n, , w wh hi il le e l lo ow w c co os st t n no ot t. .
S Sp pe ec ci if fi ic c E En ne er rg gy y C Co on ns su um mp pt ti io on n ( (S SE EC C) ) o of f P PT T
S SE ER RM MA AN NI I S ST TE EE EL L d de ec cr re ea as se e a ab bo ou ut t 0 0. .7 7 k kW Wh h/ /t to on n a a
m mo on nt th h a af ft te er r d do oi in ng g e en ne er rg gy y c co on ns se er rv va at ti io on n. . T Th he e m me et th ho od ds s
u us se ed d a ar re e r re ed du uc ci in ng g d de em ma an nd d c ca ap pa ac ci it ty y ( (c co on nn ne ec ct te ed d l lo oa ad d) )
a an nd d i im mp pr ro ov vi in ng g p po ow we er r f fa ac ct to or r u us si in ng g c ca ap pa ac ci it to or r b ba an nk k. .
T To ot ta al l e en ne er rg gy y s sa av vi in ng g i is s 6 62 25 5 k kW Wh h a a m mo on nt th h, , a an nd d a av ve er ra ag ge e
s sa av vi in ng g i in n 1 10 0 y ye ea ar rs s o ob bt ta ai in ne ed d i is s R Rp p. . 3 3. .0 03 36 6. .5 51 12 2, ,- - a a
m mo on nt th h. .
P PT T S SU UR RA AB BA AY YA A W WI IR RE E d de ec cr re ea as se e S SE EC C a ab bo ou ut t 7 7, ,3 34 4
k kW Wh h/ /t to on n i in n a a m mo on nt th h a af ft te er r i im mp pr ro ov vi in ng g p po ow we er r f fa ac ct to or r a an nd d
r re ed du uc ci in ng g h ha ar rm mo on ni ic c d di is st to or rt ti io on n a as s s st te ep ps s i in n c co on ns se er rv ve er r
e el le ec ct tr ri ic ca al l e en ne er rg gy y. . T To ot ta al l e en ne er rg gy y s sa av vi in ng g i is s 1 18 8. .3 37 78 8. .7 73 39 9
k kW Wh h a a m mo on nt th h w wi it th h t th he e a av ve er ra ag ge e s sa av vi in ng g g ge et t i in n 1 10 0 y ye ea ar rs s
i is s R Rp p. . 4 4. .1 14 42 2. .4 49 97 7, ,- - a am mo on nt th h
REFERENCES
[1] Centre for Renewable Energy Sources, Energy Audit
Guide, Part B : Systems Retrofit for Energy Efficiency,
European Commission, Directorate General for
Employment and Social Affairs, European Social Fund,
Athens, 2000
[2] Dugan, Roger C, et al, Electrical Power Systems
Quality, McGraw Hill, New York, 2003
[3] Fuller, Sieglinde K, and Petersen, Stephen R, Life-
Cycle Costing Manual, for The Federal Energy
Management Program, NIST Handbook 135, US
Departement of Commerce, Washington, 1995
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-06 1/6
Energy Conservation On Hotel Building
(Case Study in Bandung)
Rudy Setiabudy, Aji Nur Widyanto
Electrical Engineering Department, University of Indonesia
Depok 16424, Indonesia
Email : rudy@ee.ui.ac.id ; widyanto@ee.ui.ac.id
ABSTRACT-Energy conservation interpreted as
exploiting energy the efficient, rational and
effective without lessening energy needed,
productivity and comfortable. Energy
conservation can be done by energy audit. The
scope of an energy audit, the complexity of
calculations, and the level of economic evaluation
are all issues that may be handled differently by
each individual auditor and should be defined
prior to beginning any audit activities. An energy
audit can be simply defined as a process to
evaluate where a building uses energy, and identify
opportunities to reduce consumption. energy
consumption in hotels is among the highest in the
non-residential building sector in terms of absolute
values (for example, 280 kWh/m
2
in Greece, 420
kWh/m
2
in France). In this case study, there are
two opportunity of thrift able to be done by this
Hotel, installation filter harmonic and degradation
of energy with pay back period equal to 6 months.
The total opportunity can save the cost per month
equal to Rp 18.221.200,-
key words : Energy, Audit, Conservation
I. INTRODUCTION
ourisms relationship with the environment is
complex and may involve many activities that can
have adverse environmental effects, especially
related to greenhouse gas emissions and climate
change. However, tourism also has the potential to
raise awareness about environmental values and can
serve as a tool to finance the protection of natural
areas and increase their economic importance. The
two sides of this relationship can be seen in the many
municipalities of the Mediterranean basin that are, to
a large extent, dependent on tourism and therefore
also dependent on the quality of their natural
resources to attract visitors. In these communities,
energy consumption in hotels is among the highest in
the non-residential building sector in terms of
absolute values (for example, 280 kWh/m
2
in Greece,
420 kWh/m
2
in France).
In line with the increasing of the development
followed with growth of economics of Indonesia,
hence directly will bring the big challenge for
supplying energy, more than else development
followed with improvement of usage energy with
infinitude. Especially in town - metropolis, at this
time a lot of woke up by the office block, the
shopping centre and also hotel require energy big
enough. Governmental therefore have more than two
decade ago, blazed the way Konservasi Energi
Nasional Programs as indivisible shares from
Kebijakan Energi Nasional. Energy conservation
interpreted as exploiting energy the efficient, rational
and effective without lessening energy needed,
productivity and comfortable.
II. BASIC THEORY
Energy audits can mean different things to
different individuals. The scope of an energy audit,
the complexity of calculations, and the level of
economic evaluation are all issues that may be
handled differently by each individual auditor and
should be defined prior to beginning any audit
activities. An energy audit can be simply defined as a
process to evaluate where a building uses energy, and
identify opportunities to reduce consumption. There is
a direct relationship to the cost of the audit, how much
data will be collected and analyzed, and the number of
conservation opportunities identified. Thus, a first
distinction is made between cost of the audit which
determines the type of audit to be performed. The
second distinction is made between the type of
facility. For example, a building audit may emphasize
the building envelope, lighting, heating, and
ventilation requirements.
An organized approach to auditing will help us collect
useful information and reduce the amount of time
spent evaluating your facility. By splitting the audit
process into three distinct components, pre-site work,
the site visit, and post-site work, it becomes easier to
allocate our time for each step and leads to a more
comprehensive and useful audit report. The following
sections describe the tasks associated with each step
of the audit process.
Pre-Site Work
Pre-site work is important in getting to know basic
aspects of the building. This preparation will help
ensure the most effective use of our on-site time and
minimize disruptions to building personnel. A
thorough pre-site review will also reduce the time
T
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-06 2/6
required to complete the on-site portion of the audit.
The pre-site review of building systems and operation
should generate a list of specific questions and issues
to be discussed during the actual visit to the facility.
Pre-site Tasks
1. Collect and review two years of utility
energy data.
2. Obtain mechanical, architectural, and
electrical drawings and specifications for the
original building as well as for any additions
or remodeling work that may have been
done.
3. Draw a simple floor plan of the building on
8-1/2 x 11 or 11 x 17 inch paper.
4. Calculate the gross square footage using
outside building dimensions multiplied by
the number of stories.
5. Use audit data forms to collect, organize and
document all pertinent building and
equipment data
6. Develop a building profile narrative that
includes age, occupancy, description, and
existing conditions of architectural,
mechanical, and electrical systems.
7. Calculate the Energy Use Index (EUI) in
Btu/sqft/year and compare it with EUIs of
similar building.
The Site Visit
With pre-site work completed, we should have a basic
understanding of the building and its systems. The site
visit will be spent inspecting actual systems and
answering specific questions from your pre-site
review. Plan to spend at least a full day on-site for
each building. The amount of time required will vary
depending on the completeness of the pre-site
information collected, the complexity of the building
and systems, and the need for testing of equipment.
Small buildings may take less time. Larger buildings
can take two days or more. Here are some steps to
help you conduct an effective audit:
1. Have all necessary tools available on site.
2. Prior to touring the facility, sit down with the
building manager to review energy
consumption profiles and discuss aspects of
the facility we aren't able to see such as
occupancy schedules, operation and
maintenance practices, and future plans that
may have an impact on energy consumption.
3. Confirm the floor plan on our drawing to the
actual building and note major changes.
4. Fill out the audit data sheets.
5. Look at the systems relating to the ECMs
and O&Ms on your preliminary list.
6. Take pictures as you walk through the
building.
Post-Site Work
Post-site work is a necessary and important step to
ensure the audit will be a useful planning tool. The
auditor needs to evaluate the information gathered
during the site visit, research possible conservation
opportunities, organize the audit into a comprehensive
report, and make recommendations on mechanical,
structural, operational and maintenance
improvements. Post-site work includes the following
steps:
1. Immediately after the audit, review and
clarify your notes.
2. Review and revise your proposed ECM and
O&M lists.
3. Process your photos and paste or import
pictures on 8-1 /2 x 11 inch pages.
4. Organize all charts, graphs, building
descriptions, audit data sheets, notes and
photos into a 3 ring binder.
Electrical System Distribution Audit
The inefficient operation of electrical distribution
systems stems mainly from a low power factor. Power
factor correction is cost-effective when utility
penalties are imposed. Low power factors can be
improved with power factor correction devices and
high-efficiency motors. Additional energy can be
saved by installing energy-efficient transformers and
replacing existing motors with small and/or higher
efficiency motors, or by installing variable-speed
motor drives.
a. Power Factor
The total power requirement of a load is
made up of two components, namely, the
resistive part and the reactive part. The
resistive portion of a load can not be added
directly to the reactive component since it is
essentially ninety degrees out of phase with
the other. The pure resistive power is known
as the watt, while the reactive power is
referred to as the reactive volt amperes. To
compute the total volt ampere load it is
necessary to analyze the power triangle
indicated below:
P(Watt)
QL (VAr)
Qc(VAr)
S = P + QL
Fig 1. Power Triangle
P P
cos
S
2 2
Q P
= =
+
(1)
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-06 3/6
Q = Volt Amperes Reactive (VAR)
P = Watts
S = Volt Amperes (VA)
Cos = Power Factor
Amperes Factor
b. Harmonic
Distortion wave (IEC 702-07-43) is
transformation a signal which not be
intended and generally not realize with
fundamental signal reference 50 / 60 Hz. The
prime of Wave distortion for example
harmonic, DC component and inter
harmonic. Harmonic component or habit
referred as harmonic is wave having original
number fold frequency to fundamental
frequency. Equation of Frequency harmonic
is:
h = n x F Hz (2)
with :
h: -n order harmonic frequency.
F: fundamental frequency of system (50Hz/6
Hz)
n: harmonic order.
Fig. 2 Forming process of Harmonic Distortion
wave
Wave which is distortion formed by
result of merger fundamental wave with a
few harmonic wave which also in form of
sinusoidal. Value of harmonic amplitude of
wave is some percentage of amplitude
fundamental waves.
Harmonic spectrum distribution all
harmonic component amplitude as function
from its harmonic order and illustrated to
use histogram. At Fig. 2, can be seen
example of harmonic spectrum. It can be
said that spectrum represent current
comparison or harmonic frequency voltage
to fundamental frequency voltage or current.
Fig. 3. Harmonic Spectrum
Besides can be explained graph, wave
which distortion by harmonic earn also
explained mathematically use Fourier
analysis. Periodic wave which not in form of
sinusoidal can be expressed as amount of
expressed fundamental frequency harmonic
series with the following Fourier analysis
[4]:
0 0 0
1
( ) ( cos sin )
n n
n
f t a a n t b n t
=
= + +
(3)
with :
0
0
1
( )
T
a f t dt
T
=
(4)
= fundamental value of f(t) to one
period
that is from 0 till T
0
0
2
( )cos
T
n
a f t n tdt
T
=
(5)
= 2 x average value of f(t) cos n to one
period wave
0
0
2
( )sin
T
n
b f t n tdt
T
=
(6)
= 2 x average value of f (t) sin n to one
period wave
n = harmonic index
The Fourier equation can be used to
break distortion wave which have become
waving harmonic wave and basis. This
matter becomes basis in analyzing harmonic
at electric power system. On the contrary if
we know voltage amplitude value or current
every order, hence can be obtained by total
current or voltage as follows:
1
1
2
rms
V V =
and
1
1
2
rms
I I =
(7)
max
max
2
2 2 2 2
1 2 3
1
1 1
...
2 2
h
rms h h
h
V V V V V V
=
= = + + + +
(8)
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-06 4/6
max
max
2
2 2 2 2
1 2 3
1
1 1
...
2 2
h
rms h h
h
I I I I I I
=
= = + + + +
(9)
Total Harmonic Distortion or THD
expressing distortion level generated by all
harmonic components and defined as
follows:
2
2
1
n
n
n
M
THD
M
=
=
=
(10)
with :
THD = Total Harmonic Distortion
M
n
= RMS value of current or voltage
n
orde
M
1
= RMS value of current or voltage
at
fundamental frequency
III. Measurement and Data
This hotel subscribe to electrics from PLN
equal to 1385 kVA (tariff faction B3). From tables
and curve consume the electrics during one year,
average of usage kWh is around 429.180 kWh per
month. At figure.1 seen the usage energy of electricity
in April is highest usage around 383.046 kWh and the
lowest usage in March around 265.038 kWh. For
lowest billing happened in March (210 million rupiah)
with usage energy of electrics equal to 265.038 kWh.
Table 1. electric bill
Pemakaian Pemakaian
AUGUST 2,005 309,400 73,360 382,760 252,722,665
SEPTEMBER 2,005 305,320 72,460 377,780 259,218,405
OCTOBER 2,005 278,530 64,616 343,146 231,867,505
NOVEMBER 2,005 233,770 54,718 288,488 211,957,895
DECEMBER 2,005 298,000 68,748 366,748 291,198,465
JANUARY 2,006 291,822 65,250 357,072 284,333,365
FEBRUARY 2,006 254,590 59,334 313,924 255,776,595
MARCH 2,006 214,024 51,014 265,038 210,742,695
APRIL 2,006 311,342 71,704 383,046 258,970,710
MAY 2,006 283,096 64,280 347,376 266,925,980
JUNI 2,006 297,348 69,440 366,788 283,009,865
JULI 2,006 262,924 62,176 325,100 271,952,190
BULAN
PEMBAYARAN
TAHUN BIAYA TOTAL
LWBP WBP
Konsumsi Energi
-
50,000
100,000
150,000
200,000
250,000
300,000
350,000
400,000
450,000
Agst. 05 Sept. 05 Okt. 05 Nov. 05 Des. 05 Jan. 06 Febr. 06 Mar. 06 Aprl. 06 Mei. 06 Jun. 06 Jul. 06
Bulan
E
n
e
r
g
i (k
W
h
)
-
50,000,000
100,000,000
150,000,000
200,000,000
250,000,000
300,000,000
350,000,000
B
ia
y
a
(R
p
)
Energi Biaya
Fig. 3 Graphic of electric energy and bill
Measurement consumes electrics energy and
harmonic levels have been done at electrics panel in
company making of bearing. Then, the data has been
recorded by 9625 Power Measurement copied to
computer personal and accessed to use 9625 Power
Measurement Support Software.
Obtained by Measurement data for as follows:
a. Harmonic of Voltage
Value of Total Harmonic Distortion (THD) of
Voltage at the panel as follows:
Table 2. Value of Total Harmonic Distortion (THD) of
Voltage
VOLTAGE HARMONIC
U
1
(%) U
2
(%) U
3
(%)
Average 1.82 1.95 1.95
Maximum 2.35 2.59 2.36
Minimum 1.51 1.61 1.63
By using maximum Voltage THD equal to 3 %,
seen that voltage harmonic level at electrics panel
have in tolerance range.
Hereunder represent each phase of distortion
voltage wave at the electrics panel which have
harmonic. Fig. 4 hereunder represents waveform
noted by measuring instrument.
Fig. 4. voltage wave
2. Current and Level of Harmonic Current
Value of Total Harmonic Distortion (THD) of
Current at the panel as follows:
Table 3. Value of Total Harmonic Distortion (THD) of
Current
CURRENT HARMONIC
I
1
(%) I
2
(%) I
3
(%)
Average 14.04 14.04 15.15
Maximum 20.51 20.51 24.40
Minimum 8.33 8.33 8.26
By using maximum Current THD equal to 10
%, seen that current harmonic level have passed
tolerance range. Fig. 5 represents each phase of
harmonic distortion current wave which have and
noted by measuring instrument.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-06 5/6
Fig. 5. Current wave
Compared to voltage wave, influence harmonic
distortion current wave stronger destroying. This
matter can be seen from level of harmonic current (%
THD) larger than level of harmonic voltage (% THD).
3. Power Factor
Value of Power Factor at the panel as follows:
Table 4. Value of Power Factor
PF
1
PF
2
PF
3
Average 0.99 0.99 0.98
Maximum 1.00 0.99 0.99
Minimum 1.00 0.98 0.97
By using minimum power factor equal to 0,85,
seen that ower level have in tolerance range.
4. Total Power
Fig. 7 represents of total power which have and
noted by measuring instrument.
Konsumsi Energi
-
100.000
200.000
300.000
400.000
500.000
600.000
700.000
0
:0
0
1
:0
0
2
:0
0
3
:0
0
4
:0
0
5
:0
0
6
:0
0
7
:0
0
8
:0
0
9
:0
0
1
0
:0
0
1
1
:0
0
1
2
:0
0
1
3
:0
0
1
4
:0
0
1
5
:0
0
1
6
:0
0
1
7
:0
0
1
8
:0
0
1
9
:0
0
2
0
:0
0
2
1
:0
0
2
2
:0
0
2
3
:0
0
Waktu
E
n
e
r
g
i
(
W
)
Harian Hari Kerja Hari Libur
Fig. 6. Total power curve
From graphic, seen the result measurement
of electrics load PLN, the maximal load almost come
near 770,5 kW with mean power factor equal to 0,98
hence used energy at this Hotel almost come near
786,22 kVA or to load contract PLN (1385 kVA)
have weared the load equal to 56,77 %.
IV. Analyze and Solution
Identify energy conservation based on specification
data of electrics installation and result data of
measurement electric energy characteristic, hence
some opportunities of thrift able to be done as follows
:
a. Installation Filter Harmonik
From data and measurement, maximum
level of current harmonic have passed the
tolerance (< 10 %). To overcome the
mentioned needed by usage filter harmonic.
This Harmonic current height will cause
some losses at equipments operation,
overheating, neutral overloading, degradation
lifetime equipments and improvement of
consumption kwh ( current) so that to solve
can be done with installation filter harmonic,
thereby deductible current harmonic till
under 5 %.
Based on calculation, Cost effective for
external time of peak load (
LWBP)kWh/month = Rp.8.136.000,- and
Cost effective for time of peak load (
WBP)kWh/month = Rp. 2.275.200,-. So the
total Cost effective kWh/month = Rp.
8.136.000 + Rp. 2.275.200 = Rp.
10.411.200,- or equal Rp. 124.934.400,- per
annum.
b. Degradation subscribe from PLN
Demand factor is the number show
comparison between maximum electricities
(peak load) with electricity attached
(Subscribed kVA). Pursuant to record
measurement which have been done in
September 2006, we get calculation of peak
load/burden operate for is 786,22 kVA. With
subscribe capacity from PLN equal to 1385
kVA hence level of Demand Factor to month
of September 2006 is
Demand Factor =
786 kVA
1385 kVA
= 56,77 %
From data energy consumption (kWh a
yearlong, seen the highest consumption kwh
happened in April, is so that needed to do
approach of peak load during the month of
September to month April as follows :
383046
786, 72 797, 69
377780
x = =
kVA
Relate at calculation of peak load
(kVA), hence to determine customer capacity
(KVA) can be countable by adding reserves
KVA equal to 20 % from highest load
(kVA), so that = 120 % x 797,69 kVA =
957,22 kVA
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-06 6/6
Degradation subscribe from 1385
kVA become 1110 kVA, hence can be done
by the cost-saving kVA equal to :
Cost effective kVA/month
=( 1385 1110) x cost/kVA
= 275 x Rp. 28.400
= Rp. 7.810.000
Cost effective kVA/annum
= 12 x Rp. 7.810.000
= Rp. 93.720.000
Degradation subscribe from 1385 kVA
become 1110 kVA need an investment
around Rp 53.695.000,- so the simple pay
back period :
53.695.000
7.810.000
= =
6,875 month 7 month
V. Conclusion
1. Demand factor of electrics in this Hotel only
56,77 %
2. Peak load oh this hotel equal to 786,22 kVA
3. There are two opportunity of thrift able to
be done by this Hotel, installation filter
harmonic and degradation of energy with
pay back period equal to 6 months.
4. The total opportunity can save the cost per
month equal to Rp 18.221.200,-
References
1. Fan Wang. 2001. on Power Quality and
Protection. Gteborg : Chalmers University of
Technology
2. Roger C. Dugan..[et al]. 2002. Electrical Power
Systems Quality. New York : McGraw-Hill
3. Theraja B.L et.al. 1997. A Text Book of
Electrical Technology. New Delhi : S. Chand &
Company Ltd
4. Thumann, Albert. William J. Handbook of
energy audits, Younger6th ed.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-07 1/2
Indonesia Energy Revolution Scenario and Projection
using MESAP/Planet Simulation Model
Rinaldy Dalimi
1
, Bayu Indrawan
2
, Sven Teske
3
1
Electrical Engineering Department , University of Indonesia
2
Engineering Center, University of Indonesia
Tel. 08194831323, fax. 021-7863506, email : bayu@bayuu.web.id
3
Climate & Energy Unit Greenpeace International
Abstract Two scenarios up to the year 2050 are
outlined in this report. The Reference Scenario is
based on the reference scenario published by the
International Energy Agency in World Energy
Outlook 2006, extrapolated forward from 2030
have been calculated using the MESAP/PlaNet
simulation model. The Energy Revolution Scenario
has a target for the reduction of worldwide per
capita carbon dioxide emissions to less than 1.3
tonnes per year by 2050.
Indonesia Energy Revolution Scenario
describes a development pathway which turns the
present situation into a sustainable energy supply.
Exploitation of the existing large energy efficiency
potential will reduce primary energy demand from
the current 6,900 PJ/a (2004) to 9,500 PJ/a in 2050.
This compares with a demand of 16,000 PJ/a in the
Reference Scenario.
The electricity sector will have the strongest
growth in renewable energy utilisation. By 2050,
more than 60% of electricity will be produced
from renewable energy sources. A capacity of 112
GW will produce 367 TWh/a of electricity. In the
heat supply sector, the contribution of renewables
will continue to grow, reaching more than 65% by
2050. By 2050 over 40% of primary energy
demand will be covered by renewable energy
sources.
Keywords - Indonesia Energy Revolution Scenario,
Reference Scenario, Alternative Scenario,
MESAP/PlaNet simulation model, Renewable Energy.
I. INTRODUCTION
he climate change imperative demands nothing
short of an Energy Revolution. At the core of this
revolution will be a change in the way that energy is
both produced and distributed. The five key principles
behind this shift will be to implement clean,
renewable solutions, especially through decentralised
energy systems, to respect the natural limits of the
environment, to phase out dirty, unsustainable energy
sources, to create equity in the use of resources and to
decouple growth from the consumption of fossil fuels.
Decentralised energy systems, where power and heat
are produced close to the point of final use - avoiding
the current waste of energy during conversion and
distribution - will be central to the Energy Revolution,
as will the need to provide electricity to the two
billion people around the world to whom access is
presently denied.
II. BASIC THEORY
The development of future global energy demand is
determined by three key factors:
Population development: the number of
people consuming energy or using energy
services.
Economic development, for which Gross
Domestic Product (GDP) is the most
commonly used indicator. In general, an
increase in GDP triggers an increase in energy
demand.
Energy intensity: how much energy is
required to produce a unit of GDP.
Both the Reference and Energy Revolution scenarios
are based on the same projections of population and
economic development. The future development of
energy intensity, however, differs between the two,
taking into account the measures to increase energy
efficiency under the Energy Revolution Scenario.
III. EXPERIMENTAL RESULTS
Combining the projections on population
development, GDP growth and energy intensity
results in future development pathways for final
energy demand in Indonesia. These are shown in
Figure below for both the Reference and Energy
Revolution Scenarios. Under the Reference Scenario,
total energy demand will more than double from the
current 4,500 PJ/a to 11,300 PJ/a in 2050. In the
Energy Revolution Scenario, we expect a much
T
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-07 2/2
slower increase to 6,500 PJ/a in 2050, which is about
45% more than today, but reducing by 40% the
projected consumption under the Reference Scenario.
The accelerated increase of energy efficiency, which
is a crucial prerequisite for achieving a sufficiently
large share of renewable energy sources in energy
supply, is beneficial not only for the environment but
also from an economic point of view. Taking into
account the full service life, in most cases the
implementation of energy efficiency measures saves
costs compared to additional energy supply. The
mobilisation of energy saving potential leads directly
to a reduction in costs. A dedicated energy efficiency
strategy therefore helps to compensate in part for the
additional costs required during the market
introduction phase of renewable energy sources.
IV. CONCLUSIONS
The Energy Revolution scenario shows a pathway to a
more sustainable future for Indonesias energy sector.
Its main features are improvements in energy
efficiency and a higher share of renewable energy.
This report demonstrates that renewable energy is not
a dream for the future it is real, mature and can be
deployed on a large scale. Decades of technological
progress have seen renewable energy technologies
such as wind turbines, solar photovoltaic panels,
biomass power plants and solar thermal collectors
move steadily into the mainstream.
REFERENCES
[1] Baur, Jorg, MESAP - Methodological Concept of
PlaNet, Institut fr Energiewirtschaft und Rationelle
Energieanwendung, Stuttgart, 1998
[2] Sven Teske, Energy Revolution A Sustainable
World Energy Outlook, Greenpeace International,
2007.
3a4737f339
Reference Scenario
0
2.000
4.000
6.000
8.000
10.000
12.000
2004 2010 2020 2030 2040 2050
Transport Other Sectors Industry
P
J
/
a
0
2.000
4.000
6.000
8.000
10.000
12.000
2004 2010 2020 2030 2040 2050
Transport Other Sectors Industry
P
J
/
a
Energy Revolution Scenario
P
J
/
a
0
2.000
4.000
6.000
8.000
10.000
12.000
2004 2010 2020 2030 2040 2050
Transport Other Sectors Industry
P
J
/
a
0
2.000
4.000
6.000
8.000
10.000
12.000
2004 2010 2020 2030 2040 2050
Transport Other Sectors Industry
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-08 1/4
Ocean Wave Power Plant With Air Pressure
Massus Subekti*, Vivian Karim Ladesi
Fac. of Engineering, University of Indonesia,
Telp. 08881547426, email: massus.subekti@ui.ac.id; massussubekti@yahoo.com
Fac. of Oceaninc, Sepuluh November Institut of Technology
Telp. 08563100083, email: karimladesi@yahoo.com
Abstrak- This Paper study about device a power plant
of sea wave with pressure air, this generating
advantage among other things source energi
abundance especially the coastal area and archipelago
area, high closeness energy and the potential area
which unlimited. done wave observation of area to be
used for reference, then design the effective model to
convert sea wave energy become observation data
input so that will be got turbine energy which couple
with generator
Result of research indicate that by using
calculation Boyansi, level of force float (F
1
) at
secondary ponton have diameter 30 cm is 0.14 kN or
140 N. With l1 = 4 m and l2 = 2m, hence got force at
cylinder equal to 280 N. with Wide of Output Surface
7,85.10
-5
m2, Force of cylinder output is 0.27 N.
Correct territorial water for the use model this
conversion is coast of between 10 m until 100 m from
coast lip which adapted by contour sea floor of wave
and coast break condition which possible happened,
for this case is coastal area Lampung. Level of Energy
which convertion very hinge from high of waves that
happened, so that in the end will have an in with the
level of yielded electricity.
Keywords Ocean Wave, Power Plant
I. INTRODUCTION
n this time, most of energy used come from fossil fuel,
that is oil fuel, gas, and coal, used fossil fuel can
destroying the environment wich induce for green house
beside that, fossil fuel is nonrenewable and unsustainable
source energy category.
This problems increase complex when a lot of society
not benefit from electric current, especially on purilies,
isle, seaboard and not yet been reached by netork of
electric are, because the location is far from electric
power station. This condition of couse can pursue
economic growth in this are. For this problem, alternative
souce energy which reach to this area.
One of alternative souce energy the society neds is
exploiting sea wave energy. The profit from this anergy
is enough for supplying high density of energy, unlimited
of potential area, neednt fuel, no pollution, renewable
enrgy, more produces of energy, and cheap cost.
II. BASIC THEORY
Ocean Wave
Wave is one form of transition or switch simple
energy wich that energy received from original source
kinds of sea wave based on generator:
1. Bodies moving on surface make a wave at sea
surface with low periode from low energy
2. Swells on the sea surface from blown wind in the sea
surface
3. Tsunami from disruption of earthguake
4. The tide causing flooding from month and sun
gravitation field
One of its, the important and use in this research is
swells
Many types of wave in the world and difficult to
describe by matematis because unlinear, tree dimension
and its for wich random. But, there are some approach
can be used, among the others is Airy theory, stokes,
mich, Gersner, konidal, and Single, with each other
having different definition.
SWELL (linear wave) is a certain of wave as relativie
have great length and small height. SWELL is represent
wave was energy. Although high wave (H) more small
than long wave (), energy from high SWELL relative.
For the comparison between height wave to length
wave(H/) 1/50.
[1]
.
I
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-08 2/4
Figure 1. Sket of wave
Float Building Theory
The observed rom based low of static of float building
theory, because of that the se water fluid not get out from
Archimedes principle where afloat buiding gotten
immersed to water will experience force comparable with
water volume removed, this mean weight of object float
at this fluid same as water volume remove. level of force
that happened at object following equqtion:
F = g V
[2]
(1)
Where : F = Force float (N)
g = Gravitation
V = Volume
= Specific gravity fluid
While, g = 10,06 kN/m
3
atau 1,026 t/m
3
(2)
So that the level of force will influence by volume of
object float.
At This research, the secunder pontoon ude is circle
form, if the assumed is pontoon have diameter (d) so
pontoon volume is
Vponton = 4/3 (d/2)
3
(3)
Wind Energy Convert
Large of energy was transferred by wind energy to
turbine rotor shaft depend on air density, wide turbine
rotor are and air speed, and can be formulated:
E = m v
2 [2]
(4)
dimana : E = energy (joule)
m = mass of air (kg)
v = wind speed (m/detik)
Sum up mass which through of place by wide A m
2
and move with speed m/detik, can be formulated:
m = A v q
[2]
(5)
where : A =Wide (m
2
)
q = air density (kg/m
3
)
large of power which tranfered by turbine heve
efficiency can be formulated:
[3]
P = E (per et of time)
= A q v
3
= r
2
q v
3
(kW)
[3]
(6)
dimana : P = Forve (kW)
= efficiency of turbine rotor
r
2
= wide area of turbien rotor (m
2
)
r = radius of turbien rotor (m)
q = air density (kg/m
3
)
v = wind speed (m/det)
III. EXPERIMENTAL RESULTS
Diagram Alur
Research step path in the form of diagram can be
described by following :
Figure 2. Groove diagram of the research
Detailedly step of research:
1. Literature Study
Collecting circumstantial and more complete
literature-literatur about theory and method which
can be utilized in research
2. Models Scheme of conversion Energy
Determining of model convert energy which is
assuming effective and efficient
3. Determining size measure model
Determining wich size measure model have
making to be conducted a calculation
4. Determination result of and the compatible
location
Simulation/ Case study And place which is suited
for application
S
H/
C
H
h
x
z
SWL : Still water Level
Literature Study
START
Scheme Model.
Size measure Model
Result of.
FINISH
Yes
No
?
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-08 3/4
L = 3m
P = 5 m
T = 2m
Pon
ton
Sek
Model and the Size Measure Convert
1. Convert Energy sketch
Figure 3 Convert energy sketch ( seen from front)
2. Diagram Block
Figure 4. Block of diagram of generating device
3. Principle
Energy yielded by secondary ponton from sea wave
channelled to pistone through lever. Piston will
depress air come into air tube 1& 2. the high
pressured air used to turn around turbine which have
been tied on a generator. then the Air thrown
4. Primary Ponton
Figure 5. Primary ponton dimension
Figure 6. High of Primary Ponton from sea floor
Figure 7. Place Arrangement of primary Ponton
( seen from the)
5. Secondary ponton & Activator Lever
Figure 8. Dimension of activator lever.
By using calculation Boyansi (Eq.1), level of force
float (F
1
) at secondary ponton have diameter 30 cm (V
=1.4137x 10
-5
is 0.14 kN or 140 N. With l1 = 4 m and l2
= 2m, hence got force at cylinder equal to 280 N
Primary
Ponton
7 m
F
1
F
2
4
2
piston
G
Generator
Air turbin
Air
save 2
Air Save 1
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-08 4/4
F
2
6. Cylinder
Figure 9. Dimension of cylinder
Diameter of piston (
piston
) = 32 cm
Wide of piston surface = 0,08 m
2
Diameter of cylinder output = 1 cm
Wide of Output Surface = 7,85.10
-5
m
2
Force of cylinder output (F
3
) = 0.27 N
Long of Cylinder = 0,5 m
Air volume @ cylinder = 0,04 m
3
Air volume 20 cylinder = 0,8 m
3
6. Air Save
Diameter = 0,25 m
Long = 1 m
Volume Air Save 1 &2 = 0,098 m
3
Diameter output = 5 cm
Surface output = 19.63 cm
2
= 1.96.10
-3
m
2
IV. CONCLUSIONS
1. Ocean Wave Power Plant enough to be used
especially in archipelago region which not yet been
touched by an electrics network
2. Correct territorial water for the use model this
conversion is coastal area with distance 10 m until
100 m from coast lip which is adapted for contour of
sea wave floor condition and coast break which
possible.
3. Level of energy converted very depended from high
of coastal wave that happened, so that in the end will
have an in with the level of yielded electricity
REFERENCES
[1]. Triatmodjo, Bambang, Teknik Pantai, Beta Offset,
Yogyakarata, 1999
[2]. Giancoli, Daoglas, Fisika 1, 5
th
edition, Erlanggga,
Jakarta, 1998
[3]. Abdul Kadir, Energi, Sumber daya, Inovasi, Tenaga
listrik dan Potensi ekonomi, Edisi Kedua, UI Press,
Jakarta, 1995
[4]. Hans Jk, Ocean Energy Reecovery, Honolulu.
Hawai 1989
[5]. Lewis, Edwar V.. Principles Of Navar Architecture,
Printed in The USA, Jersey City,1988
[6]. www.oceanpowertechnologies.com/pdf/senate_hearing_pa
per.pdf
[7]. www.Wavegen.co.uk
: 32cm
30cm
:1cm
F
3
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-09 1/4
T
The Time for Indonesia to Optimize Geothermal
Energy
Oleh: Massus Subekti*
* Fac. of Engineering University of Indonesia
Telp. 021-4712137, email: massus.subekti@ui.ac.id; massussubekti@yahoo.com
Abstract- world oil price almost 100 US dollar per
barel progressively push world society to call
away the attention new source energy utilize to
anticipate period rare of petroleum. One of source
energy which nowadays start to become world
attention is geothermal, besides big enough
potential and the environmental friendliness,
convert goethermal become electrics own
efesiency which high enough.
exploiting of Goethermal directly have reached
27.825 MWt in 71 state in the year 2004. While for
the power station have reached 8.200 MWe in 21
state in the world. among other things the United
States of America, French, Germany, and
Indonesian.
Indonesia represent state with biggest potency
geothemal in the world that is 40 % or equal to
27.000 MWe, but only 807 MWe or equal 3%
have been exploited for power station. Some
constraint like weakness still the infrastructure,
low price sell, weakness clarity punish,
governance bureaucracy which the condition still
KKN, subsidize fuel oil and also weakness
awareness socialize, making growth PLTP
Indonesia walk very tardy.
Some step which ought to be done among other
things gift of incentive to development PLTP,
covering: reduction PBB, area retribution, PPH,
PPN, Iease import and the licensing amenity.
Revise legislation about geothermal in order to
give clarity punish to all investor, reduction step
by step subsidize to BBM in order to the price sell
of geothemal more kompetitif, facility of research
activity, seminar and discussion in around
technological development of PLTP in Indonesia,
and which do not less important planning
inwroughtly and continual in development of
national energy in order to there continuity in sill
of energy of primary in Indonesia
Keywords Geothermal, Indonesia
I. INTRODUCTION
The price of petroleum which keeps
increasing has forced the world citizen to start
the change their attention towards the latest energy
resources in order to anticipate the extinction of
petroleum era. One of the energy resources which
become the attention of the world is geothermal since
not only the energy that is produced can be conversed
into electric energy, but its potential is also sufficient
and friendly for the environment.
Several countries who have already applied
geothermal energy such as the United States,
England, France, Italy, Sweden, Swiss, Germany,
New Zealand, Australia, and Japan
II. BASIC THEORY
According to etymological term, Geothermal is
originated from two words, Geo which means earth
and thermal which means heat. Therefore
geothermal means the heat that comes from the earth.
According to the regulation in UU No. 27/ 2003
about geothermal, it is stated that geothermal is the
resource of heat energy contained in hot water, steam
and rocks along with the mineral substances and
other gases which genetically cannot be separated in
geothermal system and its function needs a mining
process.
The movements of the earth layers which collide
with each other causes the radioactive process in
earths layers depth and make a resulting magma
formation with reservoir more than 2000
o
C. Every
year rain and snow trickles absorb into the earth
layers, and are held in rocks layers which are already
exposed to heat waves and magma. These rocks
layers are called geothermal reservoir which range
between 200
o
300
o
C. The water cycle which
happen every year cause the reservoir rocks layers as
the place to create geothermal energy which can be
produced repeatedly in a very long time.
In the earths skin, sometimes the water stream
can be very close with hot rocks where the
temperature can reach until 148
o
C. The water does
not turn into steam because there is no contact with
air. When the hot water comes up into the earth
surface because of a space or a crack on the earths
skin, the hot water will come up, which is usually
called hot spring. This hot spring is commonly used
for hot spring pools, and many of these become a
tourism object.
[1]
III. EXPERIMENTAL RESULTS
A. The History of Geothermal Energy in
Indonesia
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-09 2/4
The exploration of geothermal in Indonesia was
started in 1920, Netherland built 5 shallow wells with
depth of +66 meters in West Java. Out of 5 wells
which were built, there is only 1 well left which still
produce superheated steam, that is the locomotive
well in Lumajang. In 1930, the inventory of
geothermal was once stopped because of economic
depression. Then, in 1974 Indonesia started its
geothermal exploration again.
[2]
B. The Potency of Indonesian Geothermal
Indonesia is a country with the largest geothermal
potential in the world as much as 40% from the world
reserve which equals with 27.000 MWe. This
potential is spread especially in Sumatra, Java, Bali,
Sulawesi, Nusa Tenggara Barat, and Nusa Tenggara
Timur. From that potential, only 807 MWe or + 3%
which is already exploited for geothermal power
plant. (See Table 1)
Table 1. The Geothermal Potential in Several
Countries
[3]
Exploitation
No. Country Potential
(MWe) (%)
1 Iceland 5800 202 3.5
2 USA 22990 2534 11
3 Indonesia 27000 807 3
4 Philippines 4335 1931 44.5
5 Japan 20000 535.25 2.7
6 Mexico 6000 953 15
7 New Zealand 3650 435 11.99
Based on the location of Indonesia geothermal
potential, West Java is at the highest rank with 40
locations and followed by NTT with 18 locations,
Aceh with 17 locations, and North Sumatra, West
Sumatra, and North Sulawesi with 16 locations in
each. (Table 2)
Table 2. The Locations of Geothermal Potential in
Indonesia
[4]
No Province sum No Province sum
1 Aceh 17 14 East Java 11
2 North Sumatara 16 15 Bali 5
3 West Sumatra 16 16 NTB 3
4 Riau 1 17 NTT 18
5 Jambi 8 18 Nort Sulawesi 5
6 Bengkulu 4 19 Gorontalo 2
7 Bangka Belitung 3 20 Central Sulws 14
8 South Sumatra 6 21 South Sulws 16
9 Lampung 13 22 SouthEast Sul 13
10 Banten 5 23 Nort Maluku 9
11 West Jawa 40 24 Maluku 6
12 Central Jawa 14 25 Papua 2
13 Yogyakarta 1 26 West Kalmt 3
Total Lokasi 251
Source: The Directorate General of Minerals, Coal, and
Geothermal, ESDM
Based on the category of geothermal potential, the
total amount of Indonesia geothermal resources is
14.080,5 MWe, with 9.467,5 MWe of which is
speculative and the rest of 4.613 MWe is hypothetic
resource. Whereas the total reserved potential is
13.060 MWe. Java is the island with the largest
geothermal potential with estimated reserve reaching
for 1.837 MWe with installed capacities of 785
MWe. (Table 3)
Table 3. Indonesia Geothermal Energy Potential
[4]
Resource(M
We)
Reserve (MWe)
Location
Specul
ative
Hypo
thetic
Estim
ated
Feasi
ble
Verifi
ed
Capac
ities
Sumatera 5,63 2.353 5.433 15 389 2
Jawa 2.362,5 1.591 2,86 603 1.837 785
Bali-Nusa
Tenggara
175 427 871 - 14 -
Sulawesi 925 125 721 110 65 20
Maluku 275 117 142 - - -
Kaliman 50 - - - - -
Papua 50 - - - - -
9.467,5 4.613 10.027 728 2.305 Sum
251 lokasi 14.080,5 13.060
807
MWe
27.140,5
Source: The Directorate General of Minerals, Coal, and
Geothermal, ESDM
From the existed potential, there is only 2.7%
which has been exploited while 3.1% is being in the
exploration phase, 30% is still in the geological
survey phase, and the rest of 63,7% is still being in
the preliminary survey phase. (Chart 1)
[5]
Chart 1. Percentage of Indonesia Geothermal
Energy Potential
C. Indonesia Geothermal Power Plant
Indonesia Geothermal Power Plant Capacity
The exploitation of Indonesia geothermal energy
is enhancing together with the increase of electric
energy necessities. This improvement of usage is
shown from the increasing electric energy capacity
usage from geothermal energy as much as 32,25 MW
in 1982; it increased into 142,42 MW in1990;
increased into 587,5 MW in 1998 and 807 MW in
2004. (Chart 2)
[6]
Percentage of Indonesia Geothermal
Energy Potential
Ekspl oi t asi
2. 7%
Eksplorasi
3.1%
Survei Geologi
30%
Survei Awal
63.7%
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-09 3/4
Source : various source
Indonesia Geothermal Power Plant
Capacity
32.25
142.42
807
587.5
0
200
400
600
800
1000
1982 1990 1998 2004
Year
MW
Chart 2. Indonesia Geothermal Power Plant Capacity
Current Indonesia Geothermal Power Plant
At present, there are 7 Geothermal Power Plants
(GPP) which have been operating; 140 MW in
Kamojang, 330 MW at Gunung Salak, 145 MW in
Darajat, 110 MW in Wayang Sindu, 60 MW in Dieng
(Java) and 2 MW in Sibayak (North Sumatra) and 20
MW in Lahendong (North Sulawesi) (Table 4)
Table 4. Current Indonesia Geothermal Power Plant
No Region Company
Capacity
(MW)
1 Kamojang Pertamina 140
2 Gunung Salak Chevron G.Salak 330
3 Darajat Chevron Darajad 145
4 Wayang Windu Star Energy 110
5 Dieng Geodipa 60
6 Sibayak Pertamina 2
7 Lahendong Pertamina 20
T O T A L 807
Source: Department of Energy and Mineral Resources
Exploitation Plan of Geothermal Power Plant
(GPP)
At this moment there are 3 places which are still
in the phase of exploitation plan, they are; Lahendong
II GPP in North Sulawesi with 20 MW capacities
which is planned to activate in the end of 2007, and
Wayang Windu II GPP in West Java with 110 MW
capacities which is planned to operate in 2008, and
Bedugul II GPP with 175 MW capacities which
cannot be confirmed when to operate yet, due to the
contradiction from the residents of Bali itself. (table
5).
[7], [8]
Table 5. Geothermal Power Plants in Exploitation
Plan
No Name
Capacities
(MW)
Year
1 Lahendong II
[7]
(Nort Sulawesi)
20 2007
2 Wayang Windu II
[8]
(West Java)
110 2008
3 Bedugul (Bali) 175 ?
Source: various sources
Geothermal Power Plants in Exploration Phase
At present, there are 14 GPP which are still in the
exploration phase, such as; Seulawah GPP with 160
MWe and Jaboi GPP with 50 MWe in Nanggroe
Aceh Darussalam, Sekincau GPP with 238 MWe in
Lampung, Cisolok GPP with 45 MWe, Tangkuban
Perahu GPP with 220 MWe, Gunung Tampomas
GPP with 20 50 MWe in West Java, Ungaran GPP
with 50 MWe in Central Java, Ngebel Gunung Wilis
GPP with 120 MWe in East Java, Mataloko GPP
with 65 MWe in NTT
[9]
Table 6. List of Geothermal Power Plants in
Exploration Phase
[10]
No Region Company
Capa
city
(MWe)
1 Seulawah NAD 160
2 Jaboi NAD 50
3 Sekincau Lampung 238
4 Cisolok West Java 45
5 Tangkuban Perahu West Java 220
6 G. Tampomas, sumedang West Java 20-50
7 Ungaran Central Java 50
8 Ngebel, gunung Wilis East Java 120
9 Mataloko NTT 65
10 Atadei NTT 40
11 Marana south-east
Selawesi
40
12 Suwana Gorontalo 65
13 Jailolo North Maluku 75
14 Songa-Wayaua North Maluku 140
T O T A L 1358
Source: www.esdm.go.id
Indonesia Geothermal Energy Development Plan
Referring to the Blue Print of National Energy
Management 2005 2025, the government has
planned mixed energy which in 2003 the geothermal
energy will be raised from 1.4% until 3.8% from all
national energy consumption. In order to achieve that
target, the government has made the Road Map of
Indonesia Geothermal Energy Development 2004
2025 by enhancing the capacity of GPP from year to
year. In 2008, it is aimed that the capacities will be
installed for 2000 MW whereas in 2012, 2016, and
2020 it is aimed to reach 3442 MW, 4600 MW and
6000 MW. Therefore, in 2025, the capacity is
intended to reach for 9500 MW.
Constraints in Indonesia Geothermal
Development
A number of steps that should be taken for
Indonesia geothermal development, such as;
1. The distance between the location of geothermal
resources and the capacity load makes the need
of massive infrastructural supports is important.
2. The flaws in infrastructure, especially the
transmission system that connects Geothermal
Power Plant with capacity load.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-09 4/4
3. A significant opening investment. The
geothermal development needs enormous
investment especially in exploration phase which
affects the aspects of expense and value from the
whole project, and the decide of the outcome
steam price. Geothermal Energy Association
(GEA), the Energy Department of the United
States of America noted that making an
exploration of geothermal well with the capacity
of 55MWe will need investment which costs
US$ 5.8 million. This includes the cost as much
as US$ 432 thousand for preliminary study, US $
1.2 million for exploration, and US$ 4.2 million
for properness study.
[11]
4. The fuel subsidy that makes the selling price
becomes less competitive.
5. Very low selling price which is not comparable
with the expended investment.
6. The requirement of advanced technology, which
results in the use of geothermal exploration
technology from out of the country.
7. The weakness in legal transparency that can
create the possibility of contract violation.
8. Bureaucracy constraints. It is difficult to pass
government bureaucracy in who is not
corruption-free yet.
9. The weakness in community awareness and
government role in socializing the importance of
geothermal power plant development.
The Strategies of Geothermal Power Plant
Development in Indonesia
A number of steps that should be taken for
Indonesia geothermal development, such as;
1. The distribution of incentives toward GPP
development that include the decrease in PBB,
region retribution, PPH, PPN, Import Tax and
permission ease.
2. The revision of legalizations about geothermal
usage in order to give legal transparency to the
investors.
3. The distribution of buying guarantee after the
steam is discovered. Based on the consideration
that geothermal is not an exportable commodity,
but it can be useful merely for domestic needs.
4. The fiscal policy of free-tax policy in order to
import the operational goods (PDRI) for the
necessity of geothermal effort is necessary to be
given in order to support axisting contract
project so the production capacity can be added
in accordance with the contract commitment.
5. The periodical decrease of fuel subsidy so that
the selling price of geothermal will be more
competitive.
6. Facilitating researches, seminars and discussions
about the development of geothermal power
plant technology in Indonesia.
7. Supporting the improvement of geothermal usage
for direct significance.
8. Integrated and continuous planning in improving
national energy, consequently, there will be a
synergy in the development of primary energy in
Indonesia
IV. CONCLUSIONS
If it is viewed from the existing potential, the
factor of technological progress, the increase of
worlds oil price and the efforts to reduce the glass
house effects, the future of geothermal energy in
Indonesia is very promising. On the other hand,
considering the high cost of investment needed, it
seems that the development will not be the same as
expected. It needs not only concrete steps from the
government but also complete support from the
community so that the prospect of geothermal energy
in Indonesia is compatible with the expectation.
REFERENCES
[1]. http://www.pertamina.com/index.php?option=com_co
ntent&task= view&id=3015&Itemid=340
[2]. http://www.kompas.com/kompas-
cetak/0307/14/teropong/407920.htm
[3]. www.dim.esdm.go.id/makalah/2-8%20ITB-
Nenny%20M%20S.pdf
[4]. Sugiharto Harsoprayitno, Peluang Panas Bumi
Sebagai Sumber Energi Alternatif Dalam Penyediaan
Tenaga Listrik Nasional, Direktorat jenderal Mineral,
Batubara dan Panas Bumi, Dep ESDM
[5]. http://www.majalahtrust.com/ekonomi/sektor_riil/116
6.ph
[6]. www.dim.esdm.go.id
[7]. www.kompas.com
[8]. rafflesia.wwf.or.id.pdf
[9]. www.esdm.go.id
[10]. Blue Print Pengelolaan Energi Nasional 2005-2025
[11]. http://www.republika.co.id/koran_detail.asp?id=29903
2&kat_id=&kat_id1=&kat_id2=
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-10 1/4
Incorporating Energy Commodity Price Volatility in
Economic Cost of Supply Analysis for Electricity
Expansion plan
Emil Elestianto Dardak*
* Fac. Of Economics, International Programme, INDONUSA Esa Unggul University
Tel. (62-21)5674090 email : emil.dardak@indonusa.ac.id
Abstract Rising prices of tradable energy
commodity is a key risk in the provision of electricity
infrastructure. Indonesia is undertaking major
expansion of electricity through construction of coal-
fired power plants. The current abundance of coal
supply, including the technology to utilize low calorie
coal, is the main rationale behind the selection of
coal. At the same time, the electricity utility (PLN), is
faced with the need to pursue least cost expansion,
and is given no clear incentive to take into
consideration energy mix and security in its
expansion plan. Thus, whenever coal-fired power
plant can be pursued as the financially least cost
expansion, any other generation options, even those
promising hedge against commodity price volatility,
such as renewable energy, lose out to coal. This
paper seeks to conduct a cost comparison between
coal-fired power plant and geothermal power as a
renewable energy, in which the aspect of coal price
volatility is taken into consideration. Economic cost
of supply curve will be used to conduct the analysis.
Keywords energy, electricity, renewable energy, cost
of supply, price volatility
I. INTRODUCTION
nergy security is becoming increasingly pertinent
along with the significant hike in fossil fuel price,
particularly oil. A Presidential Decree no.5 has
mandated an energy mix with more than 10%
contribution coming from new and renewable energy.
However, we see coal dominating the immediate
expansion plan with the 10,000MW crash program. At
the same time, renewable energy has grown very little.
Among the most significant barriers to the
implementation of renewable energy is the cost of
provision. Global Environmental Facility has in the past
provided incremental cost financing to provide
incentives for development of renewable energy and to
cover such incremental cost vis--vis conventional fuel
based generation. With the ratification of Kyoto
Protocol, a global mechanism named Clean
Development Mechanism was established to reward
reduction in carbon emission achieved by substituting
less clean energy technology with cleaner technology.
Nevertheless, only 4 years away from the end of the
Kyoto Protocol in 2012, Indonesia has not been able to
utilize this mechanism to upscale its renewable
generation. Geothermal, is available in huge potential in
Indonesia, with an estimated size of 27GW. However,
only 4 percent of this potential has been developed and
only 2 projects have or will soon benefit from carbon
credits, which are Lahendong and Darajat.
Therefore, it can be argued that the provision of carbon
credit cannot enhance viability of renewable energy to a
level that makes it competitive to substitute less clean
fossil fuel technology. On the other hand, the volatility
of fossil fuel price has led many to question about the
reliability of such generation for the longer term.
This is where renewable energy is believed to possess a
comparative advantage by nature, as its price is
relatively stable. Taking geothermal as an example,
geothermal promises stable price over the long term.
However, least cost expansion planning conducted by
the state power utility in Indonesia reveals bare
minimum geothermal generation, as reflected in the
five-year Electricity Provision General Plan. The
competitiveness of renewable energy is clearly low if
reference is made to this least cost planning, however,
this contradicts the notion that renewable energy can
become competitive with its stability of price. It is
argued that the lack of quantification of such
commodity price volatility hedge benefit is the main
reason for its relative absence in the terms of large scale
contribution for future power generation plan.
It is however understandable to argue that the site-
specific characteristics of geothermal project costs have
made it difficult to incorporate in the least cost
E
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-10 2/4
expansion planning model. In addition, other factors
may have hindered the development of geothermal,
which includes the level of risk, time taken from
reconnaissance to reach commissioning, and other
possible factors that may affect the comparison between
geothermal and other generation options. Nevertheless,
it is important that quantification of the benefit of hedge
against volatility can be made to allow for such
incorporation in the expansion planning for power
utilities.
II. BASIC THEORY
A study by Wiser and Bolinger (2004) look into the
volatility of fossil fuel prices and how this enhances the
competitiveness of renewable energy with its relatively
certain cost of supply. Their study focuses on the
volatility of gas-fired power plants in the USA, and on
how to appropriately compare the levelized cost of
fixed-price RE to the levelized cost of variable-price
gas-fired generation. Wiser and Bolinger also point out
the common current practice in analytic studies and
utility planning, as previously mentioned in the case of
Indonesian power utility. The practice is to compare the
cost of these resources based on an inherently uncertain
and notoriously inaccurate fuel price forecast.
Wiser and Bolinger suggest that a more appropriate
approach be the comparison of the levelized cost of RE
to the levelized cost of gas-fired generation based on a
guaranteed price of natural gas that can be locked in
with forward, futures, or swap contracts. They test the
reliability of such contracts as reference for long term
price forecasts based on futures, forward and swap gas
prices that were priced in November 2000 November
2003. The finding shows that price forecasts are
consistently made below the forward price. Hence, to
lock in a coal price, ones must pay a premium. At the
same time, price forecasts have been inaccurate, and
often underestimated the actual price of fuel
commodity. Thus, to obtain the benefit of hedge against
price volatility, utility must refer to a forward price if
non renewable energy or fossil fuel based generation is
selected.
Figure 1 shows the comparison for contract price and
spot price for the Northern Appalachian prices between
January 2000 and September 2002. Contract prices are
generally higher than spot prices, whereby contract
prices are more stable and are priced at a premium that
range between $1 and $2 per short ton. In terms of
percentage the premium ranges from 4% to 9%, with the
exception of minor cases where spot price is higher than
contract price. If we take an average of those two
numbers, the base premium for locking in coal price is
6.5%.
Figure 1 Comparison of Contract and Spot Prices for
Coal
In case of renewable energy such as geothermal, the
cost of fuel is not applicable since the source of energy
is contained in the reservoir where it is non-tradeable
and has no fluctuations of price. There is possible
fluctuation in operations and maintenance cost due to
inflation, but since such consideration is not made in
coal power calculation, this is not going to affect the
comparison results. There are indeed make up wells
requirements to maintain steam supply, but this can be
accounted for in the initial capital costing.
III. EXPERIMENTAL RESULTS
A comparison of cost for coal power and levelized cost
for renewable energy (geothermal) was made. Coal
power cannot actually be levelized as cost inevitably
increases along with long term fuel price increase.
Geothermal energy has inherent risk and thus a low case
and high case scenario were used in this calculation.
The assumptions used for calculating cost of coal power
plant: i) Capacity: 300MW, ii) Heating Value: 0.4
ton/MWh, iii) operations and maintenance: $30/KW,
iv)capital cost: $1,000/KW, iv) lifetime: 25 years, and
v) annual operation hours: 7,000 hours.
The assumptions used to calculate geothermal power
(low case) is i) Capacity: 300MW, ii) operations and
maintenance: $18/KW, iii) capital cost: $1,800/KW, iv)
lifetime: 25 years, and v) annual operation hours: 8,760
hours.
The assumptions used to calculate geothermal power
(high case) is i) Capacity: 300MW, ii) operations and
maintenance: $18/KW, iii) capital cost: $1,800/KW, iv)
lifetime: 25 years, and v) annual operation hours: 8,760
hours.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-10 3/4
In the calculation of coal power plant cost, the coal price
used is the forward coal price that is charged at an
average premium of 6.5% of the spot price. The
forecasted spot price is indexed to inflation at an
average of 6% per year. Based on this calculation, we
obtain the results where every year, coal power price
must change due to change in coal fuel price.
Using the above assumptions at the discount rate of
12%, the cost for first year when spot price of coal is
estimated at $45/ton is 4.7 cents. It is important to note
that forward price is used as volatility of price within
the year can be hedged using such contract, and this
makes the comparison with renewable energy more
robust. The cost towards the end of the lifetime at 25
years, reaches 10.6 cents due to the increase in spot coal
price at $182/ton.
If we compare this with spot price based calculation, the
first year cost would be 4.6 cents, and the cost towards
the end of the lifetime reaches 10.1 cents, which is
around 5 percent lower than the price based on forward
contracts for coal that is priced at 6.5 percent premium.
In the case of geothermal, the low case gives a levelized
cost of 3.71 cents per kWh. Levelized cost is possibly
calculated in the case of geothermal since there is no
associated fuel price increase relevant to the costing.
Variations in capital cost is based on geological
conditions, whereby depth of well and productivity are
among key factors that affect the cost of upstream
activities, in addition to the number of wells and success
ratio of drilling. The high case for geothermal gives a
levelized cost of 6.19 cents per kWh.
If geothermal discount rate is assumed at 16% to reflect
the higher risk for the project, the low case changes to
5.94 cents and the high case increases to 7.86 cents per
kWh.
If coal power plant is calculated conventionally, using
spot price at a fixed level, the levelized cost would be
4.61, in which a low case geothermal would be
competitive while a high case geothermal would not be
competitive when discount rate is set at 12%, which is
equal to coal.
A mid-range between the two cases for geothermal
would also not be competitive to this levelized cost of
coal. It is true that in the first year cost for coal in the
previous calculation of coal based on contract price,
coal is still more competitive than geothermal with the
exception of low case. But along with the increase in
coal price, the cost per kWh increases and makes
geothermal more competitive than coal.
When the discount rate is set at 16%, a conventional
calculation of coal would make both geothermal power
not competitive. If we compare both cases, in the case
of 12% discount rate, a high case geothermal would
only be viable when the project reaches its 11
th
year
when the coal power price reaches 6.2 cents. In the case
of 16% discount rate, a low case geothermal would only
be viable when the project reaches its 10
th
year when
coal power price reaches 6 cents, and the high case
geothermal would only be viable when the project
reaches 18
th
year, when coal power price reaches 8
cents.
The challenge now is in finding the levelized cost for
coal power plant. Since the price cannot be levelized,
there needs to be a benchmark for cost over the lifetime
of the project. To use the first year figure would
underestimate the cost per kWh, while using the final
year figure would overestimate since such price would
only be valid for the final year of operation.
As a solution, the benchmark price is calculated based
on the average per kWh price over the lifetime of the
project. The price is also normalized with the discount
rate to reflect the present value of higher power price
that reflects higher purchasing power increase along
with economic growth as indexed by inflation. Inflation
is assumed at an average of 6 percent per annum.
Interestingly, if we use inflation to normalize future
price, the cost of coal power becomes lower as the
increase in price is less than the discounted rate. Thus,
on the basis of future paying ability increasing
proportionate to inflation, an increase in coal price may
be acceptable and the ascending price structure may
allow for cheaper initial power price that may fit with
the existing purchasing ability.
The price of 1.06 cents in the 25
th
year actually becomes
very low when present value is calculated at 6%
discount rate, whereby the normalized price is 2.5 cents.
Thus, when the benchmark price is calculated at the
average per kWh price over the lifetime of the project,
the cost is at 3.25 cents.
IV. CONCLUSIONS
Renewable energy provides the benefit of energy
security and hedge against price volatility. Price
volatility can be contained through locking in of price
using forward contract instruments. Such contract for
coal would create premium at an average of 6.5% of the
spot price. Thus, comparison between renewable energy
and coal price must be made using this contract price,
particularly since utility seek to avoid volatility in
ensuring security of energy supply.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-10 4/4
Conventional method of assuming constant price and
trying to arrive at a levelized cost for coal fired power
plant is certainly underestimating the true cost of
provision over the lifetime of the project. Thus,
comparing renewable energy with coal becomes
difficult and biased towards coal that is currently priced
lower in the market.
The use of inflated price to reflect long term price
increase has fundamentally changed the results.
Conventional levelized cost would give 4.61 cents per
kWh, while a levelized cost would give 4.7 cents in the
first year going upwards to more than 10 cents in the
end of the 25
th
year.
Competitiveness of geothermal increases if coal price
increase is taken into account. However, if coal price
increase is normalized with inflation figures to account
for increasing ability to pay, the increase caused by coal
price can be offset by the increase in purchasing power.
This notion requires careful examination and may
undermine the impact in relation to environment and
potential scarcity due to variability of uses other than
power generation. The key finding of this analysis
would remain the fact that geothermal power promises
hedge, which can significantly alter the competitiveness
of coal.
REFERENCES
[1] Wiser, R. and Bolinger, M. (2004), The Value of
Renewable Energy as Hedge against Fuel Price
Risk, World Renewable Energy Congress
Proceeding, Colorado
[2] Globalcoal.com, Futures in Coal
[3] Platts Coal Database
EFFECTS OF CHIMNEY DEPTH AND DOWNJET
HEIGHT IN A COAL BRIQUETTE STOVE ON
THE CO EMISSION AND IGNITION TIME
Dijan Supramono*, Yulianto Sulistyo Nugroho**, and Dian Nurlita Kusuma*
*Department of Chemical Engineering, Faculty of Engineering, University of Indonesia
**Department of Mechanical Engineering, Faculty of Engineering, University of Indonesia
Depok 16424, West Java, INDONESIA
Phone: 021-7863516, email : dsupramo@che.ui.edu
Abstract - The utilisation of Indonesian coal as fuel
has increased steadily from year to year in response
to more expensive oil. Concurrently, the utilisation
of coal in briquette shapes has also increased,
especially by small and medium industries and low-
income households.
However, their increased utilisation encounters
some disadvantages such as long ignition time
(ignition delay) and high pollutant emissions. In
terms of the CO emission, it has been found that the
current briquette stoves still have high CO emission
in the order of magnitude 100 ppm, which could
reach 1000 ppm. These figures are well above the
threshold emission value of 25 ppm for human
health.
Effects of the chimney depth of the coal briquette
stove and downjet height installed inside the
chimney on the CO emission have been investigated.
Chimney is a zone above the briquette bed in the
coal briquette stove. The depth of the chimney may
affect the residence time of combustion gases above
the briquette bed and longer chimney zone is
hypothesed to allow CO to have sufficient time to
convert to CO
2
. The downjet height is measured
from the mouth of the coal briquette stove to the
surface of briquette bed surface. The downjet above
the coal briquette bed creates a recirculation above
the briquette bed which retains combustion gases for
some time in the chimney zone and enables CO to
have sufficient time to convert to CO
2
. At the same
time, unreacted hydrocarbons have also longer
residence time to convert to CO
2
. The conditions
which accommodate the conversion of CO to CO
2
are encouraged which gives two benefits, lower CO
emission from the briquette stoves and higher coal
heat release from the briquette. Exothermic reaction
from carbon in coal to CO
2
is approximately three
times of that from coal to CO.
In this investigation, data of combustion
temperature and CO emission above the coal
briquette bed versus time were collected in a fire
calorimeter during the combustion of 0.8 kg coal
briquettes starting from their ignition. The depth of
chimney and the height of the downjet were varied
respectively from 10 cm to 30 cm and from 5 cm to
20 cm. The air for combustion flowed naturally,
which is also called natural draft, from the bottom
side of the stove upwards.
The analysis of the data found that the deeper
chimney gives shorter ignition time and higher
maximum temperature. However, the deeper
chimney experienced steeper temperature reduction
after the maximum temperature was achieved. When
the maximum temperature was achieved, the CO
concentration dropped. It was followed by the
increase of CO concentration after the combustion
gas temperature dropped. CO concentration
dropped again after it reached its maximum value.
Averagely the CO concentration is higher in the
deeper chimney. In terms of the effect of downjet
height, it was found that the higher downjet
increases ignition time and maximum temperature,
but reduces the average CO concentration in the
combustion gases. In conclusion, the deeper
chimney and shorter downjet benefit to the shorter
ignition time and adversely increase the CO
concentration in the combustion gases.
.
Keywords: Coal briquettes , Stove, Chimney, Downjet
I. INTRODUCTION
Statistical data published by Ministry of Energy and
Mineral Resources showed that coal briquette
production was predicted to increase twice in the next
four years [1]. Although the utilisation of the coal
briquettes has good prospect, the high CO and
hydrocarbon emissions and long ignition time are still
major problems. Some preliminary experiment on the
related topic in the researchers laboratory found that
the current briquette stoves still have high CO emission
which could reach 700 ppm [2]. These figures are well
beyond the threshold CO emission in working areas
stipulated by Minister of Workforce in 1997, i.e. 25
ppm. Human exposed to such a high CO emission for
long time will have health problems because gas CO can
bond haemoglobin of the blood which hinders the
transport of oxygen in the blood.
The effort carried out by Tekmira Bandung to
reduce CO emission was by creating proper ventilation
in the room where briquette stove using exhaust fan. By
doing so, the gas of the product of coal briquette
combustion is expelled out of the room and fresh air is
drawn towards the room so that the CO content in the
air drops approaching the threshold value [3]. It is well
known that the ventilation system in the kitchens in
most of Indonesian households is rarely equipped with
exhaust fan. Without the fan, the convection of CO to
the exterior air is slow and this gas can remain inside the
kitchens for some time. Therefore, the introduction of a
method/methods to improve the design of coal briquette
stove so that the design itself controls the emission of
CO and hydrocarbons. This research is concerned with
how to reduce the CO gas emission which leads to
achieving or approaching the threshold value of CO
emission.
Emission CO originates from oxidation reactions of
volatile matter as well as carbon. In order to obtain
highly complete reaction towards CO
2
formation 3
requirements should be fulfilled, i.e. sufficient residence
time for conversion of CO to CO
2
, sufficient oxygen to
complete oxidation reaction, and high temperature to
improve reaction kinetics. The first requirement should
be considered in the briquette combustion because the
highest temperature of the briquette material only
reaches about 700
o
C as a consequence of the heat
absorption of the briquette combustion by high heat-
capacity briquette material [3]. This temperature is
much lower than that achieved by liquid or gas fuel
combustion, which could reach 1200
o
C. These low
temperatures are unfavourable for the conversion of CO
to CO
2
which requires high temperatures (more than
1000
o
C) [4]. In order to offset this disadvantage, coal
briquette combustion requires much longer residence
time to obtain high completion of oxidation reaction of
CO to CO
2
, which is the controlling reaction of coal
briquette combustion. The second requirement is related
to the sufficient contact between hydrocarbons of
volatile matter and CO gas and oxygen to react towards
the CO
2
formation. This is related to the mixing pattern
between briquettes and air. The excessive oxygen makes
the heat hardly increase the combustion gases
temperature high enough to proceed the conversion of
CO to CO
2
and incurs partial oxidation. Products of this
partial oxidation are oxygenated compounds such as -
CH
2
O, -CHO and CO gas [5]. The first and second
requirements can be enhanced if the reactions proceed at
high temperatures. Therefore, in order to reduce the
emissions of CO gas and hydrocarbons, in all zones
where the contact between combustion gases and
oxygen in coal briquette stove occurs, the contact time
is sufficiently long, the contact between oxygen and
partial oxidation products is performed well, and the gas
phase temperature is kept high.
II. EXPERIMENTAL
In this research, the first and second requirements
were to be investigated by varying of chimney depth
and downjet height (see Figures 1 and 2). Combustion
reactions wer carried out using natural draft for air
supply. Chimney is a zone above the briquette bed in the
coal briquette stove. The chimney depth is measured
from the mouth of the coal briquette stove to the surface
of briquette bed. The hypothesis of the research is that
the depth of the chimney affects the residence time of
combustion gases above the briquette bed and the
deeper chimney allows CO to have sufficient time to
convert to CO
2
.
The downjet is a jet installed above the coal
briquette bed which directs fresh air jet downwards. The
downjet above the coal briquette bed creates a
recirculation above the briquette bed which retains
combustion gases for some time in the chimney zone
and allows CO and unreacted hydrocarbons to have
sufficient time to convert to CO
2
. The downjet also
supplies additional oxygen for the reactions of CO and
unreacted hydrocarbons. The conditions which
accommodate the conversion of CO to CO
2
are
encouraged which gives two benefits, lower CO
emission from the briquette stoves and higher coal heat
release from the briquette. Exothermic reaction from
carbon in coal to CO
2
is approximately three times of
that from coal to CO.
The depth of chimney was chosen 10, 20 and 30 cm
by adjusting the height of grate and downjet height 5,
10, 15 cm by adjusting the length of the downjet nozzle
at chimney depth of 30 cm. For each value of chimney
depth or downjet height, the CO concentration in the
flue gas of the stove was measured. The velocity of the
air from the nozzle downjet was measured averagely
0.017 m/s.
The measurement of CO concentration was
performed concurrently with the measurement of
temperature of combustion gases immediately above the
briquette bed to obtain information of the ignition time.
Using this information, effects of the chimney depth and
downjet height on the ignition time, which is a measure
of how difficult the coal combustion works, were
obtained.
The equipment and material used in the experiment are
as follows:
1. Fire calorimeter
The fire calorimeter facilitates temperature and CO
concentration measurements. The calorimeter was built
with reference to ASTM standard E1354-1997 and ISO
5660. Figure 3. shows schematic design of fire
calorimeter. A thermocouple was used to measured flue
gas temperature sucked from the briquette stove.
2. Gas analyzer Quintox
A gas sampling gun was inserted into the ducting to
measure CO and residual oxygen concentrations in the
flue gas and connected to a Gas Analyzer, Quintox
KM9106. Using data logger of the gas analyser, CO and
residual oxygen concentrations can be measured at
preset sampling frequency.
3. Coal briquette
There were 2 coal briquettes used in the experiment,
ignition-promoting briquettes and cooking briquettes.
The first briquettes were used to accelerate the ignition
delay of the cooking briquette combustion. They were
prepared by blending the subbituminous coal powder
and 15% weight of ethyl acetate and 10% cooked starch
as briquette glue. The cooking briquettes were supplied
from South Sumatera. The weight ratio of the first to the
second briquettes was 1 : 10. The experiment was
initiated by burning ignition-promoting briquettes for 10
minutes. The burnt briquettes were then laid on the top
of the cooking briquettes and this was the beginning of
the cooking briquette combustion. The combustion
allowed to occur while CO and oxygen concentration
data were logged and flue gas temperatures immediate
above the briquette bed and in the calorimeter ducting
were recorded.
Figure 1. Stove with chimney zone
Figure 2. Stove with downjet installed above the briquette bed
III. RESULTS AND DISCUSSION
1. Chimney Depth
Profile of the flue gas temperature above the briquette
bed against combustion time is shown in Figure 4. At all
values of chimney depth, the profiles are similar in
which the temperature increases to reach a maximum
value. The times when the temperatures start to increase
which represents the initial time of the ignition are
different for different chimney depths. The result
presented in Figure 4 shows that the deeper the
chimney, the earlier the ignition starts and the deepest
chimney gives more abrupt increase of the flue
temperature. At other depths, the increase is more
gradual. In terms of the maximum flue gas temperature,
the deeper the chimney, the maximum temperature is
higher. The ignition is triggered by the release of
volatile matters from the briquette surface. Because the
ignition starts at different times among briquettes in the
stove in which the briquettes close to the ignition-
promoting briquettes (upper layers cooking briquettes)
ignite earlier, the ignition times described in Figure 4
combine different ignition times of the briquettes in the
stove. Briquette bed with the deepest chimney gives
better heat propagation from the ignition-promoting
briquettes to the cooking briquettes compared to others,
implied by abrupt increase of the flue gas temperature.
The earlier ignition time in the stove with deeper
chimney shows that the stove generates earlier heat due
to more oxidation reaction between oxygen and
briquettes. There is a competition between the
favourable condition of more natural draft due to
buoyancy driving force [6] and unfavourable condition
of more natural draft due to back pressure [7] in the
chimney at higher temperature. It seems that back
pressure is more dominant to determine the magnitude
of natural draft. This is implied by the effect of chimney
depth on CO concentration as shown in Figure 5.
Figure 5 shows that. CO molecules start to evolve
when the ignition starts to occur and its concentration
drops when the ignition reaches its maximum
temperature. It shows that CO is converted to CO
2
if
high temperature is available in the flue gas as shown by
corresponding temperature data in Figure 4. The
combustion reaction of coal briquettes is initiated by
fast reaction of coal to gas CO and slow reaction of CO
to CO
2
. The kinetics of the late reaction is enhanced if
high temperature prevails. The CO curve rises again
after the maximum temperature passes for some time
followed by the gradual reduction of CO concentration
as mass of briquettes also reduces. Visually, Figure 5
shows that deeper chimney gives more CO production.
This is in line with what has been mentioned in the
previous paragraph, in which deeper chimney prevents
the natural draft than shallower chimney. Consequently,
CO concentration in the flue gas produced by the deeper
chimney is more than that by shallower chimney. The
average calculation of the CO emission found that the
CO concentration average of the stove at chimney
depths of 30cm, 20cm and 10 cm respectively are 456,
287 and 235ppm. High CO concentration measured in
this experiment suggests that further investigation is
required to reduce this amount. One way to achieve this
may be by preheating the line of the downjet so that the
air exiting the downjet is hot enough to function as
additional reactant in the briquette surface and it is
expected that CO concentration can go lower.
Figure 3. Schematic design of fire calorimeter
2. Downjet Height
Figure 6 shows the effect of downjet height on the flue
gas temperature right above the briquette bed. The
graphs show that the stove with higher position of the
downjet from the surface of the briquette bed, the later
the ignition starts to occur. The velocity of air exiting
the downjet reduces when the air flows further down
away from the downjet nozzle [8]. The location of
downjet near the surface of the briquette bed benefits
the convection of the heat from the ignition-promoting
briquettes to the cooking briquette beneath the earlier
briquettes. Therefore, the stove with lower position of
the downjet gives earlier ignition. However, once
ignition occurs, the heat is absorbed by the downjet air
(quenching). The lower the position of the downjet, the
effect of absorption is stronger, so that the maximum
ignition temperature is lower. This gives consequences
on the CO emission where lower maximum ignition
temperature, which corresponds to the lower average
flue gas temperature, results in higher CO emission as
shown in Figure 7. The averages of CO emission as the
downjet height is varied 10, 15 and 20 cm respectively
are 747, 433 and 301ppm. The quenching whould have
been avoided if the downjet air had been preheated
such as by passing the line of downjet through the
combusting briquette bed.
IV. CONCLUSIONS
The experiment has been carried out using coal
briquettes to determine effect of chimney depth and
downjet height on the ignition time and CO emission.
The air supplied to the briquettes using natural draft.
The experiment gives conclusions that
1. The deeper chimney gives earlier ignition time and
higher maximum temperature. However, due to the
backpressure resulting from the higher combustion
gas temperature in the more confined high chimney,
such favourable temperature even gives adverse effect
on the CO emission.
2. The lower downjet gives earlier ignition to the coal
briquettes. However, once the ignition has
commenced, the cool air exiting the downjet at lower
position gives adverse effect on the combustion
oxidation.
3. Further investigation is needed to improve the effect
of the downjet on the CO emissions. One possible is
to preheat the air in the downjet line before ejecting
the air into the chimney zone.
Figure 4. Effect of chimney depth on combustion gas
temperature
Figure 5. Effect of chimney depth on CO emission in flue gas
Figure 6. Effect of downjet height on combustion gas
temperature
CO Emissions at Varying Downjet Heights
0
200
400
600
800
1000
1200
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Time (s)
C
O
(
p
p
m
)
Jet 10cm
Jet 15cm
Jet 20cm
Figure 7. Effect of downjet height on CO emission in the flue
gas
V. REFERENCES
[1]. DESM, Planning of production and utilisation of coal
briquettes 2006-2010, Departement of Energy and
Mineral Resorces Indonesia, Desember 2005
[2]. Wibowo, T. and Supramono, D., Effects of shape, size and
carbonisation of coal briquettes on the ignition time, heat
release and pollutant emissions, Student Final Project,
2007.
[3]. Balia, L., Research and development of coal briquettes in
Indonesia, Research Centre of Mineral and Coal
Technology, Departement of Energy and Mineral
Resorces Indonesia, 1996
[4]. Makino, A. Drag coefficient of a slowly moving carbon
particle undergoing combustion, Combustion Science and
Technology, vol. 81, pp. 169-192, 1992.
[5]. Turns, S.R., An Introduction to combustion, concepts and
applications, McGraw-Hill, 2nd edition, 2000
[6]. ASHRAE Handbook of Fundamentals, Chapter 26 by
American Society of Heating, Refrigeration and Air-
Conditioning Engineers (ASHRAE). Atlanta, GA: 2001.
[7]. B. Lewis, and G. von Elbe, Combustion, Flames and
Explosions of Gases, Academic Press, Inc, 1987.
[8]. Evans, D.G., Downjet combustion of coal at low
burning rates, Inst. Fuel Conf., Newcastle, N.S.W., 1966,
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-12 1/5
Scale Up Reactor for Bio-gasoline Production from
Crude Palm Oil
Bambang Heru Susanto, Anondho Wijanarko and Mohammad Nasikin
Department of Chemical Engineering, the University of Indonesia
Kampus UI Depok, Indonesia 16424
Abstract-The process of synthesizing bio-gasoline
from crude palm oil through cracking reaction was
conducted using a liquid phase batch reactor at
atmospheric pressure. The reaction temperatures
were between 300
o
C and 320
o
C, while the reaction
time were between 1 and 2 hours for each
temperature. In laboratory scale, at 300
o
C, H-zeolite
that was prepared from natural zeolite Klinoptilolite
type, was used with the catalyst/reactant (CPO)
weight ratio 1:75 to produce 10,3% bio-gasoline in
yield that have bio-gasoline fraction around 93.5%
and octane number of 122. Catalyst was prepared by
an ion exchange method using NH
4
NO
3
aqueous
solution. Addition of 5% bio-gasoline to gasoline
increases the gasoline performance similar to the
Pertamax.
This bio-gasoline production method was
scale up to the near commercial scale and for first
step, the dimension of the reactor was increased to
10 L. The reactor dimension scale up was done
base on hydrodynamic calculation with the
assumption of constant ratio of mixing power per
volume. For better homogenized temperature and
composition in the reactor, this scale up uses a
helical ribbon impeller in this mixing reactor, the
value of mixing time is around 10 second. For
producing biogasoline that can also be used as an
additive gasoline, the optimum temperature of this
reaction is similar to the result in laboratory scale.
This reaction produced 12.2% bio-gasoline in yield
that had octane number of 106 and gasoline
fraction around 34.3% (dominated with aromatic
and olefin product). The octane number of this
product was lower to the laboratory scale product,
and for using as an additive of gasoline to
Pertamax grade gasoline, it need an increasing
volume of additive up to 10%.
Key word: scale up, H-zeolite, bio-gasoline, crude
palm oil (CPO)
Introduction
The possibilities of producing fuels
from renewable resources such as biomass are
recently one the major issues in global research
activities (Idem et al., 1997; Twaig et al., 1999;
Katikaneni et al., 1996). These kinds of
technologies are more environmental friendly
and contribute effectively toward sustainable
development (Yusoff, 2004). Vegetable oils have
been the main subject of these researches (Twaig
et al., 1999). Vegetable oils mostly consist of
triglyceride, a hydrocarbon molecules formed
when a propane molecule bonded with three fatty
acids at each of the carbon atoms. This molecule
has a quite similar structure with the
hydrocarbons in crude oil. Therefore it is a very
potential alternative for the crude oil in
producing hydrocarbon fuels. Among the
vegetable oils, palm oil is known to be one of the
most potential alternatives because of its high
amount of production, especially in tropical
country like Indonesia and Malaysia (Yusoff,
2004).
Bio-gasoline was one of renewable
hydrocarbon fuel product from vegetable oil
such as crude palm oil (CPO), and have
characteristic and composition is similar to the
commercial gasoline and may be used as
commercial additive to increase the octane
number of premium gasoline and an attempt to
produce these bio-gasoline from palm oil
through cracking reaction like commonly done in
crude oil refinery recently have not been done.
This was one of the reasons to do this research.
This research was done to study the possibility of
applying the cracking reaction, in the process of
liquid fuel production from palm oil, using
cracking catalyst in a different reaction condition
than the common condition applied in crude oil
refineries. In refineries, the cracking reaction is
commonly done in gas phase reactor with high
pressure. In this research, the reaction was done
in liquid phase reactor with the reaction
temperature below the initial boiling point of
palm oil and in atmospheric pressure.
Some researches have been done to produce
liquid fuels from palm oil. Bhatia, et al have
proved that catalytic cracking of palm oil using
HZSM-5 catalyst at 350
o
C could produce light
products such as methane, ethane, gasoline,
kerosene, diesel oil, and BTX (benzene, toluene,
xylene) (Twaig et al., 1999). He also proved that
liquid fuels can also be produced from palm oil-
based fatty acid mixture which is the waste from
palm oil industry and also from used cooking oil
(Ooi et al., 2004a; Ooi et al., 2004b). Other
attempts with different researchers, different
methods, and different catalysts have also been
done to utilize the palm oil in producing liquid
fuels efficiently (Idem et al., 1997; Twaig et al.,
1999; Katikaneni et al., 1996; Yusoff, 2004; Ooi
et al., 2004a; Ooi et al., 2004b; Twaig et al.,
2003a; Twaig et al., 2003b; Twaig et al., 2004,
Wijanarko and Nasikin, 2007). This research was
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-12 2/5
design to produce liquid fuels from CPO by
catalytic cracking using H-Zeolite at 300
o
C for
producing the renewable fuel in gasoline range.
Material and Method
The catalyst used for this experiment
was a natural H-zeolite, that was prepared from
Indonesian Lampoong natural zeolite, mostly
consist of clinoptilolite type zeolite, by ion
exchange method. For 25 kg H-zeolite
preparation, natural zeolite was processed around
50 h at 500 rpm mixing rate using ammonium
nitrate solution (1 N) as the source of proton.
thereafter this produced NH
4
-zeolite form was
dried at 110
o
C around 4 h and then calcined at
520
o
C for 5 hours to omit NH
3
species for
produce H-Zeolite. Then, produced H-zeolite
was characterized by BET to find the catalyst
pore diameter.
The laboratory scale process of
synthesizing bio-gasoline from crude palm oil
through cracking reaction was conducted using a
liquid phase batch reactor with stirrer and
heating jacket at atmospheric pressure, the
reaction temperatures were between 300
o
C and
320
o
C, while the reaction time were between 1
and 2 hours for each temperature. In laboratory
scale, at 300
o
C, it used catalyst/reactant (CPO)
weight ratio 1:75. The reactor was also equipped
with a reflux tunnel in order not to make it over-
pressured without losing any light hydrocarbon
products formed in the reaction. The support of
the catalyst was prepared by an ion exchange
method using NH
4
NO
3
aqueous solution. The
density and boiling point of the reaction products
were measured to evaluate the reaction. The
products were also analyzed using FTIR
spectrometry and GCMS. To obtain the bio-
gasoline, sample of the reactor product was
distilled in an atmospheric distillation column.
The distillations were done twice to get light
hydrocarbons with as high concentration as
possible. An addition of 5% bio-gasoline that
was obtained from previous research result was
capable to increase gasoline performance similar
to commercial Pertamax. The schematic
illustration of this laboratory scale process was
shown in Fig. 1.
The pilot scale process of synthesizing
bio-gasoline from crude palm oil through
cracking reaction was conducted using a liquid
phase batch reactor with double helical ribbon
stirrer and heating jacket at the similar condition
to laboratory scale condition. To obtain the
products, sample of the reactor product was also
purified with the similar process of laboratory
scale process. The products were also analyzed
using FTIR and GCMS. The reactor design
illustration of this pilot scale process including
its detail sizing was shown in Fig. 2.
Result and Discussion
To evaluate the cracking of triglyceride
and to ensure whether cracking really occurred,
the FTIR spectra of one of the reactor products is
compared to the FTIR spectra of palm oil before
the reaction. To do the comparison, the
absorbance peak of C=O bond in the range of
1740-1745 cm
-1
is used as reference because the
amount of this bond did not significantly
changed during reaction (Silverstein and Webster,
Fig. 1. Schematic illustration of the experimental steps.
1997). Fig. 3 shows that relative to the reference
peak, the peak at 1161 cm
-1
in the spectra of
reactor product is dissapeared which was
identified in the spectra of palm oil (not shown).
It means there was CO group during the
reaction was cracking. The peak at 3000 cm
-1
that was dominant in the spectra of palm oil (not
shown), have a significant decreasing after
catalitic cracking, and this indicated that the
decreasing of carbon atom in CPO molecular
structrure was occured. Although at interval 800
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-12 3/5
Fig. 2. Batch Reactor Design using Helical Ribbon
Impeller Type (measurement unit in cm)
1000 cm
-1
and also at 700 cm
-1
was seem as a
little peak in FTIR result of bio-gasoline product,
it was indicated that formation of double bond
fungsional group and aromatic group during
cracking reaction was occured. The other results
in pilot scale and also result in smaler case were
same to the above explanation. To identify the
hydrocarbon content of the bio-gasoline, GC-MS
analysis was conducted. GC-MS analysis was
also used to identify the hydrocarbon content of
palm oil before the reaction for comparison. The
results of the analysis are shown in Table 1. Bio-
gasoline was one of renewable hydrocarbon fuel
product from vegetable oil such as crude palm
oil (CPO), and have characteristic and
composition is similar to the commercial
gasoline and may be used as commercial additive
to increase the octane number of premium
gasoline. The yield of cracking reaction in
producing gasoline-grade hydrocarbon is
defined as follow:
% 100 %) ( =
P
Y
Vol Yield
(1)
where P is the palm oil feed volume and Y is the
gasoline-grade hydrocarbons volume in the
reactor products.
The process of synthesizing bio-gasoline from
crude palm oil through cracking reaction was
conducted using a liquid phase batch reactor at
atmospheric pressure, the reaction temperatures
were between 300
o
C and 320
o
C, while the
reaction time were between 1 and 2 hours for
each temperature. In laboratory scale, at 300
o
C, a
H-zeolite was used with the catalyst/reactant
(CPO) weight ratio 1:75 to produce 10,3% bio-
gasoline in yield that have gasoline fraction
around 93.5% and octane number of 122. An
addition of 5% bio-gasoline that was obtained
from laboratory scale research result was capable
to increase premium gasoline performance
similar to commercial Pertamax.
Fig. 3. FTIR result of Bio-gasoline Product at 300
o
C.
This bio-gasoline production method
was developed from laboratory scale up to the
near commercial scale and for first step, the
dimension of the reactor was increased to 10 L
and it would be increase to higher volume. The
reactor dimension scale up was done base on
hydrodynamic calculation with the assumption
of ratio of mixing power per volume is constant.
For better homogenized temperature and
composition in the reactor, this scale up found
that using helical ribbon impeller in this mixing
reactor, the value of mixing time is around 10
second.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-12 4/5
Tabel 1.
GC-MS Analysis Results Comparison of the CPO and bio-gasoline.
% Berat
Scale Up 10 L
Komponen Skala Labarotarium
300
o
C 310
o
C 320
o
C
Parafin/olefin
C1 C4 6,13 9,44 5,9 4,32
C5 6,10 0,38 0,22 0,23
C6 1,76 0,41 1,29 1,13
C7 0,84 4,30 5,27 6,78
C8 52,63 1,12 11,62 7,05
C9 3,09 5,40 9,42 4.57
C10 7,25 16,87 12,33 14,77
C11 15,68 1,53 16,98 6,41
C12 - 2,21 3,73 2,15
C13 - 3,28 4,06 2,41
C14 - 2.84 4,05 4,88
C15 - 14.10 6,71 7,5
C16 1.55 23,2 5,39 21,65
C17 - 5,45 1,81 1,81
C18 4,97 4,65 2,01 4,32
C19 - 0,15 0,09 3,69
C20 - 2,61 - -
Aromatik
C6 - - 0,13 0,46
C7 - - 0,88 0,62
C8 - - 0,43 0,67
C9 - - 2,21 0,14
C10 - 0,15 1,35 0,45
C11 - 0,39 2.39 1,59
C12 - 0,44 1,21 0,55
C13 - 0,76 0,64 0,66
C14 - 0,32 - 1,19
C15 - - - -
C16 - - 0,18 -
Physical Properties
Octane Number 122,4 106 98 98
Density (g/cm
3
) 0,77 0.84 0.81 0.84
Viscosity (poise) 0,054 0.0048 0.0167 0.0359
RVP (kPa) 48,32 - - -
Distillation Temp. (
o
C) 255 280 280 280
Yields (%) 10,3 12,2 14,2 8,4
Note : Grey boxes are bio-gasoline fractions ( around 93,48; 34,28; 70,28; dan 49,42% were respectively bio-
gasoline fraction in laboratory and pilot scale that was done at 300; 310; dan 320
o
C) .
For producing an additive to premium gasoline,
the optimum temperature of this reaction is
similar to the result in laboratory scale, it was
found at 300
o
C (catalyst/reactant weight ratio
1:75 and reaction time 1 hour). This reaction
produced 12.2% bio-gasoline in yield that had
octane number of 106 and gasoline fraction
around 34.3% (dominated with aromatic and
olefin product). The octane number of this
product was lower to the laboratory scale
product, and for using as an additive of premium
gasoline to Pertamax grade gasoline, it need an
increasing volume of up to 10%.
Conclusion
Using H-zeolite catalyst, this process
had succeeded in synthesizing gasoline-grade
hydrocarbons (bio-gasoline) through cracking
reaction even in atmospheric pressure. The bio-
gasoline produced contains C
8
to C
12
and also
aromatics, which also composing the gasoline
from crude oil, with the yield and octane number
of 11.93% and 122.4 in laboratory scale and
12.2% and 106 in pilot scale at the same process
condition. Addition of 5% bio-gasoline to
gasoline was capable to increase premium
gasoline performance similar to commercial
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-12 5/5
Pertamax. However, for using as an additive
gasoline to Pertamax grade gasoline, it needs an
increasing volume of additive up to 10%.
Reference
Idem, R.O., Katikaneni, S.P.R., Bakhsi, N.N.,
1997. Catalytic Conversion of Canola oil to
Fuels and Chemicals: Role of Catalyst
Acidity, Basicity, and Shape Selectivity on
Product Distribution. Fuel Process. Tech.
51, 101 - 125
Katikaneni, S.P.R., Adjaye, J.D., Idem, R.O.,
Bakshi, N.N., 1996. Catalytic Conversion of
Canola Oil over Potassium-Impregnated
HZSM-5 Catalysts: C2-C4 Olefin
Production and Model Reaction Studies. Ind.
Eng. Chem. Res. 35, 3332 - 3338
Ooi, Y.S., Zakaria, R., Mohamed, A.R., Bhatia, S.,
2004a. Catalytic Conversion of Palm Oil-
Based Fatty Acid Mixture to Liquid Fuel.
Biomass Bioenerg. 27, 477 478
Ooi, Y.S., Zakaria, R., Mohamed, A.R., Bhatia, S.,
2004b. Catalytic Cracking of Used Palm Oil
and Palm Oil Fatty Acids Mixture for The
Production of Liquid Fuel: Kinetic
Modelling. Energ. Fuels 18, 1555
Silverstein, RM., Webster FX., 1997.
Spectrometric Identification of Organic
Compounds, 6
th
Ed. John Wiley & Sons,
Inc., New York
Twaiq, F.A., Zabidi, N.A.M., Bhatia, S., 1999.
Catalytic Conversion of Palm Oil to
Hydrocarbons: Performance of Various
Zeolite Catalysts. Ind. Eng. Chem. Res. 38,
3230 - 3232
Twaiq, F.A., Zabidi, N.A.M., Mohamed, A.R.,
Bhatia, S., 2003a. Catalytic Conversion of
Palm Oil Over Mesoporous Aluminosilicate
MCM-41 for The Production of Liquid
Hydrocarbon Fuels. Fuel Process. Tech. 84,
105 - 109
Twaiq, F.A., Mohamed, A.R., Bhatia, S., 2003b.
Liquid Hydrocarbon Fuels from Palm Oil by
Catalytic Cracking over Aluminosilicate
Mesoporous Catalysts with Various Si/Al
Ratios. Micropor. Mesopor. Matl., 64, 95 -
96
Twaiq, F.A., Mohamed, A.R., Bhatia, S., 2004.
Performance of Composite Catalysts in
Palm Oil Cracking for the Production of
Liquid Fuels and Chemicals. Fuel Process.
Tech. 85, 1283 1287
Wijanarko A. and M. Nasikin, Biogasoline
production from palm oil by cracking using
NiMo/Zeolite Catalyst, Japan-Indonesia
Bilateral Symposium on Sustainable
Engineering, Yogyakarta, Indonesia, May
14
th
, 2007
Yusoff, S., 2004. Renewable Energy from Palm
Oil Innovation on Effective Utilization of
Waste. J. Cleaner Production. 14, 87 93
ISSN 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-13 1/4
Flame Lift-up on A Bunsen Burner;
A Preliminary Study
Cokorda Prapti Mahandari
*
and I Made Kartika Dhiputra
= ln (2)
Smoke extinction coefficient, k, (m
-1
) can be
determined by a smoke meter as follows :
=
I
I
L
k
o
ln
1
(3)
Concentration of smoke can also be expressed as
percentage obscuration or smoke opacity (%), given by:
% 100
o
o
I
I I
(4)
The rate of smoke production, RSP (m
2
/s), can be
determined by
=
a
s
s
T
T
V k RSP
&
(5)
where
s
V
&
is the volume flow rate of exhaust gas (m
3
/s), k
is smoke extinction coefficient (m
-1
), T
s
is the gas
temperature in the exhaust duct (K), and T
a
is the ambient
temperature (K).
In a fire calorimeter, the volume flow rate of exhaust
gas,
s
V
&
(m
3
/s) can be estimated using the following
expression [9]:
51 , 0
3
10 5 , 2
=
p
T
T
V
s
a
s
&
(6)
Since in the fire calorimeter assessment, one also
measure the mass loss rate, thus a quantity called as
specific extinction area, SEA (m
2
/kg) can also be defined
as :
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-06 3/5
m
RSP SEA
&
1
= (7)
=
kg
m
kg
s
s
m
2 2
=
m T
T
V k SEA
a
s
s
&
&
1
= (8)
The SEA represents the smoke extinction area per
mass of fuel burned.
III. EXPERIMENTAL
3.1 Experimental apparatus
This work was carried out using a fire calorimeter or
also known as a cone calorimeter developed in
Department of Mechanical Engineering University of
Indonesia. The fire calorimeter is a small-scale
instrument to measure ignition time, heat release rate and
smoke production of wood and other building products.
The instrument consists of a heater of a truncated cone
shape and is capable of providing heat fluxes to the
specimen in the range of 0 30 kW/m
2
, a digital
electronic balance, an electric igniter, thermocouples and
a temperature control system, a flow rate measurement
apparatus, a sample holder and a ducting systems.
Calibration of heat flux as a function of heater
temperature is performed with a heat flux meter. At a
sufficient distance a gas sample is taken using a portable
gas analyzer to record O
2
, CO and CO
2
levels. A portable
smoke meter is used to measure the smoke opacity of the
exhaust gas.
The basic design and working principle of the
apparatus was based on the cone calorimeter standard -
ASTM E 1354 [7]. The schematic diagram of the fire
calorimeter was given in other publication by the author
[3].
3.2 Sample preparation
Two tropical wood samples, i.e. Kayu Jati and Kayu
Kamper are used in this work. The samples are prepared
with sides of 10 mm by 10 mm and 16 mm thicknesses.
The wood samples were prepared with an along grain
orientation (Fig. 1). All specimens are conditioned to
constant mass at room temperature and relative humidity.
The properties of the samples are presented in Table 1.
Fig. 2 The specimens are prepared in along grain orientation.
Table 1 Properties of the samples used
Specimen Density,
(kgm
-3
)
Kayu Kamper 621
Kayu Jati 761
3.3 Procedure
A typical run was initiated by supplying electric
power to the heater which is controlled by an electronic
temperature controller using a chromel-alumel
thermocouple of type K. The fan speed is controlled by an
inverter and set at 13,3 Hz to give a constant volume flow
rate. The temperature is set to a certain value to provide a
desired heat flux level. Once the temperature is stable, and
other equipments, i.e. gas analyzer, smoke meter and
digital electronic balance are ready, the sample is inserted
to sample holder. The sample is protected from initial heat
flux from the heater using a radiation shield. The
experiment and data collection are started by opening the
radiation shield. The transient data of temperatures,
sample mass loss, exhaust gas analysis, smoke opacity
measurement are recorded simultaneously through a
DAQ system connected to a personal computer. In this
study, the measurement are focused on the smoke
produced and mass loss measurements. The samples are
exposed vertically to a constant incident heat flux of 27
kW/m
2
.
IV. RESULTS AND DISCUSSION
Fig. 3 Smoke opacity (%) and mass loss rate (gr/s) of Jati
wood measured using fire calorimeter and smoke meter.
Mass loss rate and smoke opacity curves
0 200 400 600 800 1000 1200
0.00
0.02
0.04
0.06
0.08
0.10
0.12
S
m
o
k
e
o
p
a
c
i
t
y
(
%
)
Mass loss rate (g/s)
M
a
s
s
l
o
s
s
r
a
t
e
(
g
/
s
)
Time (s)
0
5
10
15
20 Kayu Jati
Smoke opacity (%)
100 mm
100 mm 10 30 mm
Incident heat flux perpendicular to grain
Along grain
100 mm
100 mm 10 30 mm
100 mm
100 mm 10 30 mm
Incident heat flux perpendicular to grain
Along grain
16 mm
100 mm
100 mm 10 30 mm
Incident heat flux perpendicular to grain
Along grain
100 mm
100 mm 10 30 mm
100 mm
100 mm 10 30 mm
Incident heat flux perpendicular to grain
Along grain
16 mm
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-06 4/5
Fig. 3 shows a plot of sample mass loss rate (g/s) and
smoke opacity (%) against time for Kayu Jati sample. It
clearly shows that the first peak of the mass loss curve
occur soon after the sample was ignited. Similar pattern
was found for Kayu Kamper sample (Fig. 4). It indicates
that both wood species have similar physical responds to
incident heating. Meanwhile, Kayu Jati has greater mass
loss rate than Kayu Kamper.
0 200 400 600 800 1000 1200
0.00
0.02
0.04
0.06
0.08
Kayu Kamper
S
m
o
k
e
o
p
a
c
i
t
y
(
%
)
Mass loss rate (g/s)
M
a
s
s
l
o
s
s
r
a
t
e
(
g
/
s
)
Time (s)
0
2
4
6
8
10
12
Smoke opacity (%)
Fig. 4 Smoke opacity (%) and mass loss rate (g/s) of Kayu
Kamper measured using fire calorimeter and smoke meter.
As shown on Figs 3 and 4, there was delay on the
smoke opacity measurement data compared to mass loss
rate curves. This is due to differences in probes locations.
The mass of the sample was measured directly below the
sample, meanwhile the smoke probe was located in the
exhaust duct about 1 m away from the sample. A single
peak was identified for both smoke opacity curves, which
is consistent with the mass loss rate patterns.
Smoke extinction area curves
According to Eq. 8, the values of smoke extinction
area, SEA (m
2
/g) for both samples were plotted on Figs. 5
and 6. Again, the mass loss curves were also given on the
same figures to show any possible correlation between the
two curves. It is observed that the Kayu Jati sample has
greater values of SEA, especially in the earlier time, as
clearly shown on Fig. 7.
On the basis of the mass loss rate which has
proportional correlation with the heat release rate [3], the
curves on Fig. 7 can be divided into two sections. The
first section is representing the induction, moisture and
volatile release phase. This phase is known as pyrolisis
conditions where the smoke is relatively darker compared
to the flaming conditions. As reported in the literature the
smoke conversion factor of wood and plastic materials
under pyrolysis conditions has greater values that during
flaming combustion [5]. The smoke produced in flaming
combustion tends to have a large content of elemental
carbon. In fact Kayu Jati has higher density value than
Kayu Kamper (Table 1).
In practice, greater value of smoke extinction area,
SEA means reduction in visibility. As we all concerns,
visibility of exit signs, doors, and windows can be of great
importance to an individual attempting to survive a fire,
and plays important role on fire safety engineering
consideration during building and plant design.
Fig. 5 Smoke extinction area (SEA) (m
2
/g) and mass loss rate
(g/s) of Kayu Jati calculated according to Eq. 8.
Fig. 6 Smoke extinction area (SEA) (m
2
/g) and mass loss rate
(g/s) of Kayu Kamper calculated according to Eq. 8.
0 200 400 600 800 1000 1200
0.00
0.02
0.04
0.06
0.08
Kayu Kamper
Mass loss rate (g/s)
M
a
s
s
l
o
s
s
r
a
t
e
(
g
/
s
)
Time (s)
0.0
0.5
1.0
1.5
2.0
2.5
S
E
A
(
m
2
/
g
)
Smoke extinction area, SEA (m
2
/g)
0 200 400 600 800 1000 1200
0.00
0.02
0.04
0.06
0.08
0.10
0.12
Mass loss rate (g/s)
M
a
s
s
l
o
s
s
r
a
t
e
(
g
/
s
)
Time (s)
0
2
4
6
8
10
12
Kayu Jati
S
E
A
(
m
2
/
g
)
Smoke extinction area, SEA (m
2
/g)
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-06 5/5
0 200 400 600 800 1000
0
2
4
6
8
10
12
Kayu Jati
Kayu Kamper
S
E
A
(
m
2
/
g
)
Time (s)
Fig. 7 Comparison of SEA values for Kayu Jati and Kayu
Kamper.
V. CONCLUSION
The results of this study show that the smoke opacity
patterns correlate with the mass loss rate of the samples
tested. The rate of smoke production in the earlier stages
of combustion, i.e. pyrolysis condition is relatively higher
than that of flaming condition. In general, the value of
smoke extinction area (SEA) of Kayu Jati is greater than
Kayu Kampers. In practice, greater value of smoke
extinction area, SEA means reduction in visibility. This is
of our concern on maintaining the visibility of exit signs,
doors, and windows during a fire accident.
This work shows that the fire calorimeter apparatus
can be used effectively in measurement of smoke
properties.
ACKNOWLEDGMENT
This work has been funded by the Ministry of
National Education Republic of Indonesia through Hibah
Bersaing XIII (2005-2007) Research Fund. Thanks to Mr.
M. Andy Fardiansyah and Mr. Candra Gupta Marihot for
preparation of the samples and the smoke opacity
measurements.
REFERENCES
[1] Quintiere, J. G., A Theoretical Basis for
Flammability Properties, Fire and Materials, 30,
175-214, 2006.
[2] Nugroho, Y.S., Soesanto, and E. Puspiartono, Effect
of Sample Orientation and Thickness on Piloted
Ignition of Timber, Proceedings of the Intl.
Conference on Fluid and Thermal Energy Conversion
2006, Jakarta, Dec. 10 14, 2006, ISSN 0854 9346.
[3] Nugroho, Y.S., Junaedi, A., and Kiswanto, G.,
Pengembangan Kalorimeter Api untuk Karakterisasi
Sifat Bakar Material, Jurnal Teknologi, 4, 294-301,
2005.
[4] Drysdale, D., An Introduction to Fire Dynamics, 2nd
Edition, John Wiley & Sons, John Wiley & Sons,
2003.
[5] Mulholland, GW, (2002), Smoke Production and
Properties, Chapter 13, The SFPE Handbook of Fire
Protection Engineering, 3rd Edition, pp. 2-258 2-
268.
[6] Wikipedia, website: http://en.wikipedia.org/wiki/
Beer-Lambert_law (seen on 21
st
November 2007).
[7] ASTM E 1354-97, Standard Test Method for Heat
and Visible Smoke Release Rates for Materials and
Products Using an Oxygen Consumption
Calorimeter, Annual Book of ASTM Standards,
American Society for Testing Materials,
Philadelphia, 1997.
[8] Myllymaki, J and Baroudi, D, Prediction of smoke
production and heat release rate by convolution
model, Nordtest Technical Report 1297-96, VTT
Building Technology, Finland, 1999.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-16 1/5
Catalytic Ozonation Of Endosulfan
In Water With Activated Carbon
Enjarlis,
a,b
Setijo Bismo,
a
Slamet,
a
Roekmijati W. Soemantojo
a
a
Department of Chemical Engineering, University of Indonesia, Depok16424 Tel. 7863515
b
Department of Chemical Engineering, Institute Technology of Indonesia, Serpong Tangerang 15311 Tel.
7561092 .fax7561092.e-mail: en_jarlis@yahoo.com
Abstract The degradation of endosulfan by catalytic
ozonation with activated carbon as catalyst was
investigated at neutral pH and different
temperatures. The aim of this work is to study the
effect of the addition of activated carbon on the
degradation rate of endosulfan, focusing on the rate
constant (k) and the activation energy (Ea).
Endosulfan was selected here as target organic
pollutant due to its included organochlorine pesticide
(OCPs) may cause serious environmental
concern.The presence of the activated carbon as
catalyst is effect slightly in enhancing ozonation rate
of endosulfanan compared to non catalytic
ozonation. The rate constant (k) of endosulfan were
11.54 x 10
-2
min
-1
at 30
o
C for catalytic ozonation, and
10.27 x 10
-2
min
-1
for non catalytic ozonation. The
activation energy for catalytic ozonation was 6.58
kcal/mol and for non catalytic ozonation was
8,631kcal/mol.
Keywords Endosulfanan, Catalytic ozonation,
Activated carbon.
I. INTRODUCTION
atalytic ozonation with activated carbon as catalyst can
be accelerated ozone decomposition to form OH
radical
, thereby increasing even further the removal efficiency
of this because OH is powerfull, non-selective chemical
oxidant, which acts very rapidly with most organic
compounds [1, 2]. It has been reported that hydroxyl
radicals were generated by combinations of ozone plus
UV radiation, UV radiation plus H
2
O
2
, UV radiation
plus Fentons reagent (photo-Fenton system) [3] and by
mixing a few milligrams of activated carbon in ozone-
containing water [4]. With such context, this study was
investigated the degradation of endosulfan in water by
catalytic ozonation at neutral pH and different
temperatures, which is focused on (i) determining
possible diffusion limitation (ii) determining the rate
constants (k) and activation energy (Ea).
Endosulfan was selected here as target organic pollutant
due to its paradigmatic endocrine discruptors included
organochlorine pesticide (OCPs) may cause serious
environmental concern and health problem in animals,
including humans [5]. According to free radical reaction
theory, OH will attack chlorinated organic compounds
by hydrogen abstraction or electron transfer. Then, the
organic radical will decompose further to chlorinated
intermediates. These intermediates are eventually
oxidized by OH to final products; organic acid and
even carbon dioxide [6].
II. BASIC THEORY
The degradation of contaminant organic in presence
of activated carbon is a combination of competing
homogeneous and heterogeneous reaction: direct
reaction of molecular ozone, and an indirect reaction
involving non selective free-radicals. Both reactions take
place in the solution bulk and on activated carbon
surface [7]. Therefore, the total degradation rate of
contaminant organic in the presence of activated carbon
can be defined as the sum of the homogeneous reaction
rate, (-r
C
)
homogen,
calculated in the absence of activated
carbon, and the heterogeneous reaction rate, (-r
C
)
heterogen
due to the presence of activated carbon []. The total
endosulfan degradation rate can be mathematically
expressed as:
r r r Hete E Homo E Total E . . .
+ =
(1)
As discussed earlier the activated carbon enhanced the
oxidation rate of contaminant by promoting the
reduction of dissolved ozone into hydroxyl radicals that
cause the endosulfan degradation, the rate of equation
given in Tabel 1. Recently, Jans and Hoigne [1] pointed
out that catalytic ozone decomposition can be
cataloqued as another Advance Processes Oxidation
(AOP) with a stoichiometric ratio (ozone
decomposed/hyroxyl radical formed) equal to that of the
other ozone involved in AOPs.
Tabel 1. Rate of equation at degradation of endosulfan
by ozonation in the presence activated carbon
N
o
With activated carbon With out activated carbon
1 (- dC
E
/dt)
hete
= f
1
(C
E
, C
O3,
C
OH
, C
Ac
)
(- dCE/dt)
homo
= f
4
(C
E
,
C
O3,
C
OH
)
C
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-16 2/5
2 (- dC
O3
/dt)
hete
= f
2
(C
E
, C
O3,
C
OH,
C
AC
)
(- dC
O3
/dt)
homo
= f
5
(C
E
,
C
O3,
C
OH
)
3 (- dC
OH
/dt)
hete
= f
3
(C
E
, C
O3,
C
OH
, C
AC
)
(- dC
OH
/dt)
homo
= f
6
(C
E
,
C
O3,
C
OH
)
III. EXPERIMENTAL RESULTS
Endosulfan (1,2,3,4,7,7-hexachlorobicyclo-2,2,1
heptene 2,3-bis-hydroxy methane-5,6 sulfite) was
obtained from Chem. Service West Chester with purity
95 %. The endosulfan solution was prepared by
deionozed water obtained from Aquatron Auto Still
Yamato Tipe W-182. The specific surface area of
activated carbon were measured using the multipoint
BET of N
2
adsorption in a Quantachrome Autosorb -6
with surface area 6.87.10
2
m
2
/g. Ozone was produced
by a RS 09805 ozone generator with maximum ozone
production capacity of 0.25 g of O
3
/h. The experimental
instrument consists of an ozone generator, a cylindrical
glass column reactor with an external jacket surrounded
and a water stream which was pumped from
thermostatic bath to maintain the temperature at the
selected value for each experiment. The dimension is of
the reactor is 450 mm high with ID 40 mm which
equiped by inlet diffuser for bubbling the gas mixture,
outlet gas, sampling port and magnetic stirrer. Once the
experiment was started, the air-ozone mixture was fed
into the flasks (KI solution) in order to determine the
ozone concentration in the gas form.
The reactor was filled with 300 ml demineralized-water
and the pH was adjusted to 7. The temperature was set
for 20, 25 and 30
o
C, until a predetermined of water was
saturated with ozone in excess continuously by injecting
ozone gas for 30 min. The process is followed by the
addition 0.5 g of activated carbon (for the catalytic
process) and 4.5 x 10
-5
mol/L endosulfan in solution.
The concentration of endosulfan and ozone presence in
the system was measured at 0, 3, 6, 10, and 15 min of
treatment. The dissolved ozone concentration in aqueous
solution was determined by iodometric methods, which
endosulfan was analyzed by Gas Chromatograph type
4C, colum silicone ov-17 3 meters, ECD (Electron
Capture Detector), Shimadzu, solvent n-hexsane and
mobile gas-phase N
2.
All experiments were carried out
in duplo the presence and the absence of activated
carbon.
The influence of variables on the degradation of
endosulfan were first assessed. Non-catalytic (blank
experiment) and catalytic experiments were conducted
for each experimental series. Since the catalytic
ozonation process involves steps of external and internal
diffusion, variables studied were: activated carbon
particle size and agitation speed. From experimental
result agitation speed equal or higher than 750 rpm and
carbon particle equal or lower 0.2 -0.3 mm did not affect
the ozonation rate. Hence, at these conditions, it can be
assumed that external mass transfer rate and internal
diffusion rate of ozone through pores did not control the
catalyst ozone decomposition [9]. Therefore, agitation
and activated carbon particle particle were kept at 750
rpm and 0.2-0.3 mm size. At these conditions, the
kinetics should only be depended on the chemical
surface ozone decomposition reaction [9]. Next, a series
of experiments were then carried out at different
temperatures to determining Ea from k.
Endosulfan degradation
Fig. 1 shows the influence of the presence of activated
carbon as catalyst for the degradation rate of endosulfan
by catalytic and non-catalytic ozonation. From that
figure, it can be seen that the degradation rate of
endosulfan increased with the presence of activated
carbon compared to the absence of activated carbon.
This is because, the presence of activated carbon can
accelerate ozone decomposition to form OH
radicals
which have a high reactivity to most organic
micropollutants [1, 2].
0,5
1,0
1,5
2,0
2,5
3,0
3,5
4,0
4,5
5,0
0 3 6 9 12 15
time, min
C
E
.
1
0
. -
5
m
o
l
/
L
Fig. 1. Endosulfanan degradation rate by ozonation
processes (x) non-catalytic (blank experiment)
()catalytic ozonation at condition: pH 7; temperature
20
o
C; agitation speed 750 min
-1
;activated carbon particle
0.2-0.3 mm C
O3
6.372 x 10
-3
O
3
mol/L; CE
o
5.0 x 10
-
5
mol/L
Influence of temperature on the degradation of
carbofuran
Fig.2 shows the effect of temperature on degradation of
endosulfan for catalytic ozonation at pH 7. From that
figure, it can be stated that the higher the temperature is,
the higher is the degradation rate of endosulfan. This
effect is due to an increase in the rate constant of the
chemical reaction. Yazgan et al [10] reported that the
change of temperature gave positive effect on the
removal rate of the endosulfan. Although increasing
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-16 3/5
temperature causes decreasing in dissolved ozone
concentration (data are not shown), a higher oxidation
rate was obtained at higher temperatures. This
phenomenon can be explained by the increasing
diffusion coefficient at higher temperature [10].
0
1
2
3
4
5
6
0 3 6 9 12 15
time, min
C
E
,
x
1
0
-
5
,
m
o
l
/
L
Fig.2 The effect of temperature on degradation of
endosulfan by catalytic ozonation at condition: pH 7;
agitation speed 750 min
-1
;activated carbon particle 0.2-
0.3 mm C
O3
6.372 x 10
-3
O
3
mol/L; C
Eo
5.0 x 10
-5
mol/L;
temperature 20
o
C, 25
o
C, 30
o
C
Kinetics and Mechanism
The observed endosulfan degradation in the presence of
activated carbon results in a combination of competing
homogeneous and heterogeneous reaction: direct
reaction of molecular ozone, and an indirect reaction
involving non selective free-radicals [7]. The
degradation kinetics of endosulfan by homogeneous
reaction in terms of molecular ozone and hydroxyl
radical that is produced by the decomposition of ozone
at neutral pH can be formulated as follows:
Degradation rate of endosulfan homogeneous reaction:
+ =
OH E
E
O E
E
o
E
C C k C C k
dt
dC
. . . .
2 3 1
hom
.
(2)
Degradation rate of endosulfan by heterogeneous
reaction:
AC OH E
E
AC O E
E
AC E
E
hete
E
C C C k C C C k C C k
dt
dC
. . . . . . .
5 3 4 3
.
+ + =
(3)
Thus, since of ozone concentration in solution in excess
and the constant amount of activated carbon, the
amount of OH and O
3
stayed constant during the
processes (i.e., at a steady state). Therefore, Eq. (1), (2)
and (3) can be rearranged to the pseudo-first-order
equation in Eq. (4) as follows:
( )
E
E
Hete
E
Homo
E
Hete
E
Homo
E
C k k
dt
d
r r
C
Total
. + = + =
(4)
with,
+ =
OH
E
O
E E
Homo
C k C k k . .
2 3 1
(5)
and ,
+ + =
OH
E
O
E
AC
E E
Hete
C k C k C k k . . .
5 3 4 3
(6)
Integrating, with initial condition C
E
= C
Eo
leads to:
t k t k k
C
C
E
Total
E
Hete
E
Homo
E
Eo
. ) ( ln = + =
(7) with, .
E
Hete
E
Homo
E
Total
k k k + =
(8)
where k
E
total
represents the apparent pseudo-first-order
total rate constant for ozonation process in presence
activated carbon. The degradation endosulfan by
homogeneous reaction non catalytic as follows:
t k
C
C
E
Homo
E
Eo
. ln =
(9)
Figure 3 and 4 describe the plot of ln C
Eo
/C
E
) versus
time (t) with slope as k
E
total
and k
E
Homo
in various
temperatures by catalytic and non catalytic oznation,
respectively. Insert of Fig. 3 and 4 shows, typical
Arrehnius plot for ln k
against 1/T to endosulfan in
catalytic and non catalytic ozonation processes,
respectively. This linear regression indicate R
2
= 0.99.
From that figure, it can be stated that the higher the
temperature is, the higher is the degradation rate. This
effect is due to an increase in the constant rate of the
chemical reaction. Similar phenomenon was reported by
many researchers [10, 11].
The degradation of endosulfan in the presence of
activated carbon slightly significantly enhancing the
degradation rate of endosulfan with k
E
total
was 11.54 x
10
-2
min
-1
and Ea 6.58 kcal/mol at 30
o
C compared to
non-catalytic ozonation with k
E
Homo
was 10.27 x 10
-
2
min
-1
and Ea 8.631 kcal/mol. This is because of the
using of activated carbon which small surface area was
(6.87.10
2
m
2
/g) compared to which was used by the
other authors. They used activated carbon with surface
area was larger than 6.87.10
2
m
2
/g [2, 10].
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-16 4/5
0
0,4
0,8
1,2
1,6
2
2,4
2,8
3,2
3,6
4
0 3 6 9 12 15
time, min
l
n
(
C
E
o
/
C
E
)
-3,0
-2,5
-2,0
-1,5
-1,0
0,00325 0,0033 0,00335 0,0034 0,00345
1/T (1/.K)
l
n
k
t
o
t
a
l
,
m
i
n
-
1
Fig 3 Constants rate of degradation endosulfan by
catalytic ozonation at condition: pH 7; agitation speed
750 min
-1
;activated carbon particle size 0.2-0.3 mm;
C
O3
6.372.10
-3
O
3
mol/L; C
Eo
5.10
-5
mol/L; temperature
20
o
C, 25
o
C, 30
o
C. Insert: Arrhenius plot for
degradation of carbofuran by catalytic.
On other hand, compared to the degradation of
carbofuran by catalytic ozonation using the same
activated carbon as catalyst we obtained a slower
degradation rate of endosulfan [12]. This is because, the
reaction of endosulfan to ozone and OH
is less reactive
compared to the reaction of carbofuran to ozone and
OH
.
In addition, the presence of activated carbon in
ozonation processes does not really act as a catalyst, but
rather as an initiator and/or a promotor for ozone
transformation into OH
(2)
indicating that the liquid film controls the mass
transfer.
III. EXPERIMENTAL RESULTS
Hollow fibers membranes used in the experiment
were made of polypropylene supplied by AKZO. The
fibers are 2.7 mm in outer diameter and 0.45 mm in
wall thickness. The hollow fibers modules were made
by potting both ends of the fibers into 1.5 cm nipple
using epoxy (Araldite). The module was connected to
external shell made from PP pipe with a PP T-joint,
and at the other end of the nipple was connected to
vacuum pump using brass reducer 1.5 cm to 0.3 cm.
The diameter of the external shell can be varied using
appropriate connection to PP T-joint where in this
study 1.6 cm shell diameter was used. There were 3
contactors of about 40-cm length and 10, 15 and 20
fibers used in the experiments to give membrane
packing fractions in the contactors of 0.2, 0.3, and 0.4,
respectively.
Schematic of experimental configuration as shown
in Fig.1 was used to measure mass transfer
performance and pressure drop of the membrane
contactors. The liquid in the reservoir, which is water
or 0.1 M NaOH solution, was initially pumped
through the membrane contactor at the shell side.
Meanwhile, CO
2
gas was sent to the contactor through
the lumen of fiber. pH and temperature of the liquid
were automatically measured every 20 s from the pH-
meter probe, and digital manometer measured liquid
pressure differences between inlet and outlet
contactor.
Figure 1. Schematic diagram of experimental set up.
The overall mass transfer coefficients from the
experiments are calculated using Equation (3),
1
*
0
*
ln
C C
C C
A
Q
K
m
L
=
(3)
where C*, C
t
and C
0
are CO
2
concentrations in
equilibrium, liquid inlet and outlet contactor, Q
L
is
liquid flow rate to the membrane contactor, A
m
is
outer membrane fibers surface area, and K
L
is the
observed overall mass transfer coefficient,
respectively. Friction factors of the contactors were
calculated using Equation (4),
2
2
l
v l
Pd
f
e
=
(4)
where f is friction factor of the module, P is liquid
pressure drop, d
e
is equivalent diameter of the
contactor, l is contactor length, is water density, and
v
L
is water velocity, respectively.
CO
2
flux through the membrane contactors, J
CO2
,
and the overall mass transfer coefficients, K
L
, from
the experimental results were plotted versus water
velocity v
l
as shown in Figure 2 and 3, respectively.
As it can be seen from Figure 2, the water can absorb
CO
2
through the membrane contactor as high as
4.5x10
-8
mole CO
2
per square meter membrane area
per hour. Meanwhile, 0.01 M NaOH solution can
absorb CO
2
through the membrane contactor as high
as 5.5x10
-3
mole CO
2
per square meter membrane
area per hour, which is 120.000 times higher than
using water as solvent.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-19 3/5
The CO
2
flux decreases with increasing contactor-
packing fraction for the same liquid velocity. The
same effect also occurred for the mass transfer
coefficient of the membrane contactors. This can be
explained as in a region of low packing fraction
contactor, transverse flow and surface renewal effect
is seem to be more influence to the mass transfer
performance rather than channeling effect, meanwhile
at the higher packing density the channeling effect is
more dominant [1]. From another point of view, the
geometrical dependencies of boundary layer profiles
might also increase the mass transfer coefficient of
contactor with lower packing density. The increase is
directly related to the boundary layer conditions,
where the profiles of boundary layers on a curved
body become thinner with an increasing degree of
curvature, as for example decreasing d
o
fiber with
other conditions constant. Therefore, based on this
theory, the mass transfer coefficient will increase with
decreasing packing fraction of the contactor,
especially at higher liquid velocities. The channeling
occurred because of the existing regions of densely-
and loosely-packed in the contactor which creates
preferential flow around fibers in the contactor
resulted mainly from uneven distribution of fibers and
flow [9]. Membrane contact area might have been
reduced in densely packed regions as fibers were
more adhered to each other, reducing the availability
of contacting surface.
1.E-08
1.E-07
1.E-06
1.E-05
1.E-04
1.E-03
1.E-02
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
V
L
(m/s)
J
C
O
2
(
m
o
l
e
/
m
2
.
s
)
N=10
N=15
N=20
N=10
N=15
N=20
Figure 2. Variation of CO
2
fluxes J
CO2
, in water (filled
symbols) and 0.01 M NaOH solution (open symbols) with
liquid velocities v
L
, in the membrane contactors
To examine the dependence of CO
2
transfer
coefficient on the process parameters using water as a
solvent, correlation are conveniently expressed in
terms of dimensionless Sherwood number, Sh,
Reynolds number, Re, and Schmidt number, Sc.
C B
Sc A Sh Re = (5)
In this study the Schmidt number, Sc, was not varied,
so the 1/3 power-dependence in the literature was
assumed [10], and Equation (5) become,
33 . 0
Re Sc A Sh
B
= (6)
1.E-07
1.E-06
1.E-05
1.E-04
1.E-03
1.E-02
0.1 1
v
L
(m/s)
K
L
(
m
/
s
)
N=10
N=15
N=20
N=10
N=15
N=20
Figure 3. Variation of mass transfer coefficients K
L
, in water
(filled symbols) and 0.01 M NaOH solution (open symbols)
with liquid velocities v
L
, in the membrane contactors
Figure 4 shows the experimental data as Sh plotted
against Re at various packing fraction contactors to
obtain the exponent for Reynolds number, B in
Equation (5) for each contactor. Furthermore,
Equation (6) can be simplified into Equation (7) to see
the dependency of Reynolds number on mass transfer
performance.
b
a Sh Re = (7)
The exponent for Reynolds number, b, was
obtained by the linear regression of Figure 4 and it
was found that the values were ranging from 0.63 to
0.65 and have the average value of b=0.64. This
value indicated that the mass transfer in the contactors
is dominated by turbulent flow [10]. The values of a
for each contactor were obtained from the slope of the
experimental Sherwood number against Re
0.64
where
these values are ranging from 0.0115 to 0.0387. To
obtain the correlation between a and packing fraction
of the contactor,, the values of a were plotted against
contactor packing fraction where the correlation
between a and packing fraction is best fitted by
Equation (8),
72 . 1
0026 . 0
= a (8)
where is packing fraction of the contactor,
respectively.
Equation (8) describes the dependency of
geometry of the hollow fiber membrane contactors
employed in this study where the mass transfer
performance is related to the packing fraction of fibers
in the contactors. An empirical correlation for mass
transfer performance in the hollow fiber membrane
contactors used in the experiments can then be
correlated in the form
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-19 4/5
64 . 0 72 . 1
Re 0026 . 0
= Sh (9)
To see the effects of contactor packing fraction on
pressure drop due to friction loss, the friction factors
of the contactor were obtained from Equation (4) and
then were plotted against Reynolds number as shown
in Figure 5. The friction factor increases with
decreasing contactor-packing density for the same
contactor diameter due to an increase in hydraulics
diameter of the contactor as expressed in Equation (4).
The dependency of friction factor on Reynolds
number for fluid flow in the contactor has general
form,
b
a f Re = (10)
where f, a and b are friction factor and empirical
constants from the experimental results, respectively.
The exponent b has a negative value as can be seen
from many literatures [11,12]. As for instance, b is 1
and -0.25 for liquid flow in the laminar region
(Re<2000) and in the turbulent region (Reynolds
number up to 10
5
), respectively [11]. It can be seen
here that the negative value of b becomes smaller as
the flow more turbulent. In this study where the range
of Reynolds number measurement in the contactor is
from 800 to 3400 has b values ranging from 0.27 to
1.01 as shown in Figure 5. As a comparison, It was
found that the friction factor data has value of b = -1.1
for Re<5 and b = -0.5 for 5<Re<100 based on the
experiment using microporous hollow fiber
membrane blood- oxygenator [12].
0.1
1.0
10.0
100 1000 10000
Re
S
h
N=10
N=15
N=20
Figure 4. Variation of Sherwood numbers Sh, with Reynolds
numbers Re, in the membrane contactors
The friction factor ratio of the contactor to the
smooth pipe is shown in Figure 6. These data reveal
that the fibers surface did not behave like a smooth
pipe within the range of velocity in the experiments.
Thus viscous sub layer did not develop due to the
movement of the fibers as quantitatively expressed in
very low viscous drag values compare to inertial drag
values. The random movement of the fibers was more
pronounced at lower velocities, and as a result, the
value of the friction factor is also higher at lower
velocities [4]. The friction factor ratio in Figure 6 is
higher for lower packing densities due to the more
movement of the fibers in low packing fraction
contactor.
0.001
0.01
0.1
1
100 1000 10000
Re
f
N = 10
N = 15
N = 20
Figure 5. Variation of friction factor f, with Reynolds
number Re, in the membrane contactors
The friction factors obtained for the membrane
contactors in this study are 4.6 to 14.6 times higher
than the theoretical value based on flow of fluids in
the smooth pipe as shown in Figure 6. As
comparison, It was reported that the friction factor of
hollow fiber blood oxygenators was about 7 times
higher than the theoretical value at Re<5 and about 31
times higher than the theoretical value at 5<Re<100
[12]. Meanwhile, other researcher reported that the
friction factor of sealed end polypropylene hollow
fiber contactors with packing densities 8.7 to 9.6 %
were about 2 times higher than the values for pipe
flow reported in the literature [4]. It can be seen here
that the friction factor ratio can vary from contactor to
contactor depend on the geometry of the contactors.
0
2
4
6
8
10
12
14
0 1000 2000 3000 4000
Re
f
R
N = 10
N = 15
N = 20
Figure 6. Variation of friction factor ratio f
R
, with Reynolds
number Re, in the membrane contactors
IV. CONCLUSIONS
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-19 5/5
Experiments have been conducted to absorb CO
2
gas using water and 0.01 M NaOH solution as
solvents through hydrophobic microporous
polypropylene hollow fiber membrane contactors of
packing fraction 0.2, 0.3 and 0.4. The water can
absorb CO
2
through the membrane contactor as high
as 4.5x10
-8
mole CO
2
per square meter membrane
area per hour. Meanwhile, the 0.01 M NaOH solution
can absorb CO
2
through the membrane contactor as
high as 5.5x10
-3
mole CO
2
per square meter
membrane area per hour, which is 120.000 times
higher than using water as solvent. The CO
2
flux
decreases with increasing contactor-packing fraction
for the same liquid velocity using both water and 0.01
M NaOH solution as solvents. The same effect also
occurred for the mass transfer coefficients of the
membrane contactors.
The mass transfer performance of the membrane
contactors using water as absorbent can be formulated
as a function of contactor geometry and liquid flow
rate. The contactor geometry is represented by
packing fraction, whilst the flow rate is represented by
Reynolds number. An empirical correlation for mass
transfer performance in the hollow fiber membrane
contactors in the range of packing fraction of 0.2 to
0.4 can be correlated in the
form
64 . 0 72 . 1
Re 0026 . 0
= Sh .
Liquid pressure drops increase with increasing
contactor-packing fraction at the same liquid flow rate
due to an increase in friction between fibers and
water. Based on the friction factor data the fibers
surface did not behave like a smooth pipe within the
range of velocity in the experiments. The friction
factors obtained for the membrane contactors in this
study are 4.6 to 14.6 times higher than the theoretical
value based on flow of fluids in the smooth pipe.
REFERENCES
[1] Lipnizki, F., and Field, R.W., Mass transfer
performance for hollow fibre modules with shell-side
axial feed flow: using an engineering approach to
develop a framework, Journal of Membrane Science,
193 (2001), 195-208.
[2] Dindore, V.Y., and Versteeg, G.F., Gasliquid mass
transfer in a cross-flow hollow fiber module:
Analytical model and experimental validation,
International Journal of Heat and Mass Transfer, 48
(16) (2005), 3352-3362.
[3] Sirkar, K.K., Other new membrane processes, in
W.S.W. Ho and K.K. Sirkar (Eds.), Membrane
Handbook, Chapman & Hall, New York, (1992), 885-
899.
[4] Ahmed, T., Semmens, M.J., and Voss, M.A., Energy
loss characteristics of parallel flow bubbleless hollow
fibre membrane aerators, Journal of Membrane
Science, 171 (2000), 87-96.
[5] Lim, S.P., Tan Xiaoyao and Li, K., Gas/vapor
separation using membranes: Effect of pressure drop in
lumen of hollow fibers, Chemical Engineering Science,
55 (4) (2000), 2641-2652.
[6] Gabelman, A., and Hwang, S.T., Hollow fiber
membrane contactors, Journal of Membrane Science,
159 (1999), 61-106.
[7] Wickramasinghe, S.R., Semmens, M.J., and Cussler,
E.L., Better hollow fibre contactors, Journal of
Membrane Science, 62 (1991), 371-388.
[8] Mavroudi, M., Kaldis, S.P., and Sakellaropoulos, G.P.,
A study of mass transfer resistance in membrane gas
liquid contacting processes, Journal of Membrane
Science, 272 (1-2) (2006), 103-115.
[9] Wu, J., and Chen, V., Shell-side mass transfer
performance of randomly packed hollow fibre
modules, Journal of Membrane Science, 172 (2000),
59-74.
[10] Costello, M.J., Fane, A.G., Hogan, P.A., and Schofield,
R.W., The effect of shell side hydrodynamics on the
performance of axial flow hollow fibre modules,
Journal of Membrane Science, 80 (1993), 1-11.
[11] Coulson, J.M., and Richardson, J.F., Chemical
Engineering, Volume 1: Fluid Flow, Heat Transfer and
Mass Transfer, Pergamon Press, Oxford, New York,
Toronto, Sydney, Paris, Frankfurt, 1978, 3rd ed., pp.
40-45.
[12] Wickramasinghe, S.R., Garcia, J.D., and Han, Binbing,
Mass and momentum transfer in hollow fibre blood
oxygenators, Journal of Membrane Science, 208
(2002), 247-256.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-20 1/6
Enzymatic and Acid Hydrolysis of Bagasse for Ethanol
Production by Simultaneous Sacharification and
Fermentation
Samsuri, M.
1,2
, Gozan, M.
1
, Nasikin, M.
1
, Prasetya, B.
2
1
Chemical Engineering Department, Faculty of Engineering,
University of Indonesia, Kampus UI, Depok 16424, Indonesia,
2
Centre of Biotechnology, Indonesian Institute of Science (LIPI),
Jl. Raya Cibinong km 48, Bogor, Jawab Barat, Indonesia
Email address: msyamsuri@ristek.go.id
Abstract-Ethanol was produced from bagasse using
enzymatic hydrolysis, acid hydrolysis and
combination of Enzymatic and acid hydrolysis in
Simultaneous Saccharfication and fermentation.
Cellulase and xylanase were used for enzymatic
hydrolysis and HCl low concentrations (1% and
0,5%) was used for acid hydrolysis.
The results showed that the highest ethanol yield
with enzymatic hydrolysis with enzyme cellulase at
pH 5, about 5,51g/L or about 11, 24% base on
original bagasse. The highest ethanol production
from bagasse by combination of enzyme cellulase
and xylanase at pH 5 about 7,63 g/L or 15,25%.
The highest ethanol yield after hydrolysis with HCl
0,5% is 3,24 g/L at incubation for 24 h. The highest
ethanol yield after hydrolysis with HCl 1% is 3,77
g/L or 7,54% at incubation for 24 h. The highest
ethanol yield after hydrolysis with combination of
cellulase, xylanase and HCl 1% about 8,22 g/L or
16,45%. This yield increasing about 2% if
compared with ethanol yield without acid addition.
If compared with acid hydrolysis by HCl 1%
increasing about 8% based on original bagasse
Key words : Bagasse, SSF, Enzymatic hydrolysis
cellulose, xylanase, Acid hydrolysis HCl
1. INTRODUCTION
Most ethanol produced in the world today is
derived from starch or sucrose (Gong et al., 1999).
Ethanol fermentations are traditionally carried out for
wine or beer production, but ethanol for transportation
and industry are a large potential market and growing
use. Ethanol belongs to clean combustion. Its oxygen
content decreases emissions in mix combustion with
gasoline. Because this ethanol is originally plant
matter, its use as fuel does not contribute to the net
accumulation of carbon dioxide in the atmosphere
(Costello and Chum, 1998). Starch is abundant in crop
materials, but expansion of ethanol production for the
purpose of automotive fuel requires feedstocks that do
not compete for food or fiber (Himmel, et al., 1996).
In the other hand, lignocellulosic material is by
far the most abundant raw material such as in
hardwood, softwood, grasses and agricultural residues.
Further potential raw materials are newsprint, office
papers, municipal solid wastes, etc. Lignocellulose is a
more complex substrate than starch. It consists of a
mixture of carbohydrate polymers (cellulose and
hemicellulose) and lignin. The carbohydrate polymers
are tightly bound to lignin mainly by hydrogen bonds
but also by some covalent bonds (Lee, 1997).
Lignocellulose contains five major sugars, which
varies with the feedstock (Hinman et al, 1989). They
are the hexose; glucose, mannose, galactose and
pentose; xylose and arabinose. Glucose is the most
abundant carbohydrate in lignocellulosic material,
xylose is the second one.
One of the sustainable alternative energy is
relatively cheap production and environment friendly
was development bio ethanol from waste residue
agriculture. It does contain many lignocelluloses like
bagasse (waste residue sugar industry). Indonesia has a
lot of biomass like bagasse. Based on Indonesian
Central Research Sugar Agriculture data, bagasse was
produced 32% from sugarcane. 60% of it, was used by
sugar industry as fuel, material for paper, fungi
industry, etc. Because of that, approximately about
40% of bagasse wasnt use. Polysaccharide of
sugarcane contain more 70%, it divide to 50%-55%
cellulose and 15%-20% hemicellulose. The lignin
contain about 20-23%, beside that it was called ash
compound (Pandey, et al, 2000).
Generaly The process for converting the
lignocellulosic material to ethanol requires: (1)
delignification to liberate cellulose and hemicellulose
from their complex with lignin; (2) depolymerization
of the carbohydrate polymer to produce free sugars;
and (3) fermentation of mixed hexose and pentose
sugar as monosaccharide to produce ethanol. The
accurate analysis during conversion process is also
important for the successful development of feasible
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-20 2/6
conversion process. Presently, it is known that among
the key processes described above, the delignification
of lignocellulosic raw materials is the rate-limiting and
most difficult task to be solved (Lee, 1997).
Hydrolysis process is converted polysaccharide to
monosaccharide. Major constituents in enzymatic
hydrolysis and acid hydrolysis from lignocellulosic
material like bagasse are glucose and xylose released
from cellulose and hemicelluloses, respectively.
Ethanol fermentation process utilizing yeast has been
well developed with glucose as a carbon and energy
source. Generally enzymatic hydrolysis using
cellulose or xylanase, and acid hydrolysis is using HCl
or H
2
SO
4
.
Simultaneous Saccharification and Fermentation
(SSF) firstly described by Takagi (Takagi, et al, 1977],
combine enzymatic hydrolysis or acid hydrolysis of
polysaccharide with simultaneous fermentation of the
sugars obtained ethanol. The advantage of this process
is cellulose or hemicellulose were converted to
monosaccharide, it rapidly ferments to ethanol and use
one reactor for entire process so that saved the cost
equipment.
This research was tried to make ethanol from
bagasse by SSF process with enzymatic hydrolysis,
acid hydrolysis and combination of enzymatic
hydrolysis and acid hydrolysis.
2. Methods
2.1 Preparation of sample
Sugarcane bagasse (Saccharum officinarum) were
milled and screened (30-60-mesh). The milled bagasse
was then air-dried to a final humidity 10% and stored
under dry conditions. Bagasse was taked from
Indonesia (from sugar factory in Lampung, Indonesia).
Preparation of material Hydrolysis
HCl 0.5% and 1% was used for acid hydrolysis,
Cellulase and Xylanase was were used for enzymatic
hydrolysis.
2.2 Growth Stock S.cerevisiae
S.cerevisiae AM12 was pre-cultured in Potato
Dextrose Agar (PDA) 2%, Agar (0.25 g), Aquades
(50ml) and incubation for 3-4 days on ambient
temperature, and then was used as yeast for SSF
process.
2.3 Preparing yeast inoculation
S.cerevisiae AM12 fresh from growing stock was
pre-cultured with 50 ml medium (glucose, 10 g/l; yeast
extract, 1.0 g/l; KH
2
PO
4
, 0,1 g/l; MgSO
4
.7H
2
O, 0,1 g/l;
and (NH4)
2
SO
4
, 0,1 g/l) in 50 ml flask. Before
inoculation, medium was autoclaved for 20 minutes,
and then incubation on ambient temperature for 24
hours with use by shaker.
2.4 SSF process
SSF medium 5.5 ml, it containing bagasse (0.25
g), nutrient medium (2.5 ml), 0.5 Na-citrate buffer (pH
4, 4.5, 5) cellulase enzyme (10 FPU = 0.016 g) and 2.5
ml yeast inoculated. Bagasse, nutrient medium and Na-
citrate buffer sterilize for 20 minutes in autoclave.
However, enzyme was added without sterilization.
Nutrient medium is containing 1.0 g/l (NH
4
)
2
PO
4
; 0.05
g/l MgSO
4
.7H
2
O and 2 g/l yeast extract. Cultivation
was took and get in to micro centrifuge tube. The clean
liquid from sample had taken in 24, 48, 72 and 96
hours and doing ethanol concentration test for ethanol
produced.
For strong acid hydrolysis using 0.5% and 1%
(v/v) concentration, chloride acid was added after
bagasse and then enzyme, yeast inoculated (2 ml), Na-
citrate buffer pH 5 (0,5 ml) and nutrient (2 ml).
2.5 Analysis
a. Analysis lignin, holocellulose and -cellulose
Lignin analysis was done by modified klason
lignin method. It is with add sulphuric acid 72% in
sample and agitate until crushed, and then autoclave in
121
0
C for 30 minutes. After that, it filtered with paper
filter, covered by aluminium foil and dry it into oven
for 1 hour weighing the weight.
Holocellulose and -cellulose analysis, it analyze
by wise method. It mixed sample with sodium chlorate,
acetate acid and aquades. Incubation with hot water in
80
0
C, chilled, filtration with aquades and it wash with
acetone. Then, solid residue was dried in oven on
105
0
C for 1-2 nights and weighing the weight.
b. Measuring ethanol concentration
Ethanol concentration was measured by Gas
Chromatography SUPELCOWAX-10 type (Supelco
Inc., 0,53 mm ID., 15 m, 0,5 mm, FID) with
temperature was maintained at 50
0
C. Before tested,
sample was took 50 l and 200 l distilled water
added.
3. Result And Discussion
3.1 Composition Bagasse Analysis
The analysis result show that lignin contain in
bagasse approximately about 24.2% from bagasse total.
-cellulose contain about 52.7% and hemicellulose
contain about 17.5% in bagasse.
0 10 20 30 40 50 60
lignin
-cellulose
Hemicellulose
Extractive chemical
composition (%)
Figure 1. Chemical composition of bagasse
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-20 3/6
Totally of hemicellulose and -cellulose that is
opportunity to convert to monosaccharide in hydrolysis
process. Actually about 69% composition of bagasse
is polysaccharide and able to convert to ethanol
production. Theoritically Theoretically if calculated
base of cellulose the maximum ethanol yield about
56,86% base -cellulose and 28,43% base on original
bagasse. If calculated base of hemicellulose maximum
ethanol yield about 56,84 % base on hemicellulose and
11,37 % base on original bagasse. Totally maximum
ethanol yiel base original bagasse about 40%.
3.2 Ethanol production from bagasse with
Enzymatic Hydrolysis
Ethanol production with enzim cellulose
Ethanol production from bagasse using enzyme
cellulase showed in figure 2. The highest ethanol yield
from experiment at pH 5, about 5,51g/L or about 11,
24% base on original bagasse. Optimum pH for
enzyme is 4.8 and yeast is between 4 and 5. So that
SSF process pH is among 4.5-5. This condition is
exactly for enzyme and yeast condition. Beside pH,
temperature is also influence in SSF process.
Temperature for growing yeast media is about 30
0
-
35
0
C (Makanjuola, et al, 1992). But the cellulolytic
enzyme act optimally about 40
0
C. So, the optimum
temperature range for SSF is 35-41
0
C. At lower
temperatures the rate of saccharification decreases
significantly.
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
0 24 48 72 96
Incubation time (h)
E
t
h
a
n
o
l
c
o
n
c
e
n
t
r
a
t
i
o
n
s
(
g
/
L
)
pH = 4
pH = 4,5
pH = 5
Figure 2. Ethanol production from bagasse with
enzyme cellulose in variation of pH
At this research, SSF temperature was used in
ambient temperature, while yield ethanol was produced
is not too high because cellulase enzyme capability was
decreased to transform cellulose to glucose.
The result also shows that ethanol can be obtained
in 24 hours almost reach optimal concentration in SSF
process. After 24 hours, ethanol concentration not too
significant increase and highest ethanol concentration
was obtained between 72 and 96 hours of incubation
time. From this research, SSF process with pH 5 was
obtained the highest ethanol concentration compare
with pH 4 and pH 4,5.
Based on literature, optimum pH for SSF process
is between pH 4.5 until pH 5 and the highest ethanol
was obtained at pH 5. These phenomena is relevant
with the literature. Based of this research for the other
experiment con was continued on pH 5.
Cellulose are usually a mixture of several
enzymes. At least three major groups of cellulases are
involved in the hydrolysis process; (1) endogluconase
which attacks regions of low crystallinity in the
cellulose fiber, creating free chain-ends; (2)
exogluconase or cellobiohyrolase which degrades the
molecule further by removing cellobiose units from the
free chain-ends; (3) -glicosidase which hydrolyzes
cellobiose to produce glucose. In addition to the three
major groups of cellulase enzymes, there are also a
number of ancillary enzymes that attack
hemicelluloses, such as glucuronidase, acetylesterase,
xylanase V-xylosidase, galactomananase, and
glucomananase (Duff and Murray, 1996). During the
enzymatic hydrolysis, cellulose is degradable by the
cellulases to reducing sugars that can be fermented by
yeast or bacteria to ethanol
Ethanol production from bagasse by combination of
enzyme cellulase and xylanase.
Lignocellulosic material especially bagasse consist
17,5% of hemicellulose. The highest composition of
hemicelluloses is xylan, that can not convert to xylose
by enzyme cellulase. Enzyme xylanase is one of
alternative enzyme able to convert hemicellulose
(xylan) become xylose and xylose able to convert to
ethanol. Ethanol production by enzim cellulase and
xylanase showed in figure 3 and 4, respectively.
The highest ethanol production from bagasse by
combination of enzyme cellulase and xylanase at pH 5
about 7,63 g/L or 15,25%. The yield of ethanol if
compared with yield of ethanol using enzyme
cellubiase without xylanase increase about 3 %. This
phenomena indicated that enzyme xylanase increasing
the ethanol yield. The highest ethanol yield at pH 5,
this phenomena indicated pH 5 is the optimum pH for
process SSF.
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
0 24 48 72 96 120
Incubation time (h)
E
t
h
a
n
o
l
c
o
n
c
e
n
t
r
a
t
i
o
n
s
(
g
r
/
L
) pH = 4
pH = 4,5
pH = 5
Figure 3. Ethanol production from bagasse by enzyme
cellulase and xylanase
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-20 4/6
0.00
2.00
4.00
6.00
8.00
10.00
12.00
14.00
16.00
0 24 48 72 96 120
Incubations time (h)
E
t
h
a
n
o
l
c
o
n
c
e
n
t
r
a
t
i
o
n
s
(
%
)
pH = 4
pH = 4,5
pH = 5
Figure 4. Percentage of ethanol production from
bagasse by enzyme cellulase and xylanase base on
original bagasse
3.3 Ethanol production from bagasse with Acid
Hydrolysis
Concentrated acids HCl have been used to
hydrolysis of lignocellulosic materials (in this research
bagasse). The highest ethanol yield after hydrolysis
with HCl 0,5% is 3,24 g/L at incubation for 24 h. The
highest ethanol yield after hydrolysis with HCl 1% is
3,77 g/L or 7,54% at incubation for 24 h. Ethanol
concentration of bagasse after SSF using hydrolysis
with HCl 0,5% and 1% is showed at figure 5,
respectively.
0.00
1.00
2.00
3.00
4.00
5.00
6.00
0 24 48 72 96 120
E
t
a
n
o
l
c
o
n
c
e
n
t
r
a
t
i
o
n
s
(
g
r
/
L
)
Incubations time (h)
HCl
0,5%
Figure 5. Ethanol production from bagasse by acid
hydrolysis (HCl 0,5% and 1%).
Based on data in figure 5, ethanol production
from bagasse by HCl hydrolysis increasing at 24 h
incubation time because HCl is powerful agents for
cellulose and hemicellulose hydrolysis (Sivers and
Zacchi, 2002). The ethanol yield is lower than with
enzyme hydrolysis because HCl was used at the lower
concentrations (only 1% and 0.5%), more higher
concentration HCl the yield of ethanol will be
increasing that is indicated ethanol yield by HCl 1%
higher than 0,5%. But unfortunately concentrated
acids are toxic, corrosive and hazardous and require
reactors that are resistant to corrosion (Sun and Cheng,
2002). In addition, the concentrated acid must be
recovered after hydrolysis to make the process
economically feasible (Sivers and Zacchi, 1995).
Dilute acid hydrolysis has been successfully
developed for pretreatment of lignocellulosic materials.
The dilute sulfuric acid pretreatment can achieve high
reaction rates and significantly improve cellulose
hydrolysis (Esteghlalian et al., 1997). At moderate
temperature, direct saccharification suffered from low
yield because of sugar decomposition. High
temperature in dilute acid treatment is favorable for
cellulose hydrolysis (Millan, 1994). Recently
developed dilute acid hydrolysis processes use less
severe conditions and achieve high xylane to xylose
conversion yields. Achieving high xylane to xylose
conversion yields necessary to achieve favorable
overall process economics because xylane accounts for
up to a third of the total carbohydrate in many
lignocellulosic materials (Hinman et al., 1992). There
are primarily two types of dilute acid pretreatment
processes: high temperature (T greater than 160
o
C),
continuous-flow process for low solids loading (5-10%
weight of substrate/weight of reaction mixture)
(Brennan et al., 1986; Converese et al., 1989), and low
temperature (T less than 160
o
C), batch process for high
solids loading (10-40%) (Esteghlalian et al., 1997).
Although dilute acid pretreatment can significantly
improve the cellulose hydrolysis, its cost is usually
higher than some physico-chemical pretreatment
processes such as steam explosion. A neutralization of
pH is necessary for the down stream enzymatic
hydrolysis of fermentation process (Sun and Cheng.,
2002).
3.4 Ethanol production from bagasse with
combination of Acid Hydrolysis
This research was done by combination of cnzyme
cellulase, xylanase and additional strong acid with low
concentration in SSF process. Additional strong acid
was done together with enzyme. The strong acid was
used is chloride acid (HCl) with concentration 0.5%
and 1% (v/v). Additional strong acid will be affected in
hydrolysis process of cellulose to glucose and can be to
increase ethanol yield.
The results show that the highest ethanol yield
after hydrolysis with combination of cellulase,
xylanase and HCl 1% about 8,22 g/L or 16,45%. This
yield increasing about 2% if compared with ethanol
yield without acid addition. If compared with acid
hydrolysis by HCl 1% increasing about 8% based on
original bagasse. Ethanol production from bagasse
using SSF with combination of enzymatic hydrolysis
(Cellulase and cellubiase) and acid hydrolysis (HCl 1%
and 0,5%) showed at figure 6, respectively.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-20 5/6
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
0 24 48 72 96 120
Incubations time (h)
E
t
h
a
n
o
l
c
o
n
c
e
n
t
r
a
t
i
o
n
s
(
g
/
L
)
cellulase & xylanase
pH=5
"Cellulase, Xylanase
and HCl 0.5 %"
Cellulase, xylanase
& HCl 1 %"
Figure 6. Ethanol production from bagasse by enzyme
cellulase, xylanase and added HCl low concentrations.
Base on this result, if the acid concentration more
highly, then the hydrolysis capability will more
increase. In acid hydrolysis, the degree of
polymerisation cellulose will be decreased
significantly. It was caused by H
+
concentration
(proton) in strong acid. If the concentration strong acid
more high, then H
+
concentration will be more high
too. Hydrolysis by acid can be able to breakdown -1,
4-glycosidic chain in cellulose. The reaction will start
from proton in acid which rapidly react with glycosidic
oxygen which is connect to two sugar units transform
to conjugation acid. The breakdown of C-O chain and
broken conjugation acid will be transform to
carbonium cyclic ion. After addition water, sugar and
proton will be free (Carrasco, et al, 1994).
Hemicellulose is easier to hydrolysis rather than
cellulose with acid hydrolysis. Hemicellulose also
easier to depolymerisation in high temperature at acid
condition (Palmqvist, et al, 1996) and hemicellulose
sugar such as xylose, arabinose can be free with
hemicellulose hydrolysis. Hemicellulose hydrolysis in
acid condition is possible to form 2-furaldehyde
(furfural) form pentose sugar, it mainly from xylose
and form 5-hydroxymethil-2-furaldehyde (HMF) from
hexoses sugar. Furfural compound and HMF can
delay yeast work because it makes lag phase in ethanol
production and decrease ethanol productivity, but there
is no effect to ethanol was produced. S.cerevisae cans
metabolism furfural and HMF, in order that inhibitor
effect will be disappeared. It caused by membrane
permeability of S.cerevisae.
4. Conclusion
Based on the result of research was obtained, and
then the conclusion:
1. Ethanol was production from bagasse using
enzyme cellulase, xylanase and HCl in low
concentrations.
2. The highest ethanol yield with enzymatic
hydrolysis with enzyme cellulase at pH 5, about
5,51g/L or about 11, 24% base on original
bagasse.
3. The highest ethanol production from bagasse by
combination of enzyme cellulase and xylanase
at pH 5 about 7,63 g/L or 15,25%.
4. The highest ethanol yield after hydrolysis with
HCl 0,5% is 3,24 g/L at incubation for 24 h.
The highest ethanol yield after hydrolysis with
HCl 1% is 3,77 g/L or 7,54% at incubation for
24 h.
5. The highest ethanol yield after hydrolysis with
combination of cellulase, xylanase and HCl 1%
about 8,22 g/L or 16,45%. This yield increasing
about 2% if compared with ethanol yield
without acid addition. If compared with acid
hydrolysis by HCl 1% increasing about 8%
based on original bagasse.
6. References
Carrasco, J.E.; Saiz, M.C.; Navarro, A.; Soriano, P.;
Saef, F. and Martinez, M. (1994) Effect of dilute
acid and steam explosion pretreatments on the
cellulose structure and kinetics of cellulosic
fraction hydrolysis by dilute acids in
lignocellulosic materials. Applied biochemistry
and biotechnology. 45/46, 23-34.
Hatakka, A. (2001) Biodegradation of lignin. In M
Hofrichter and A. Steinbchel (eds.),
Biopolymers, vol. 1. Wiley-VCH, Weinheim,
Germany. p. 129- 180.
Hofrichter, M. (2002) Review: Lignin conversion by
manganese peroxidase (MnP). Enzyme Microb.
Technol. 30:454-466.
Itoh. H., Wada. M., Honda. Y., Kuwahara. M.,
Watanabe. T., 2003. Bioorganosolve
pretreatments for simultaneous saccharification
dan fermentation of beech wood by etanolysis
dan white rot fungi. Journal of Biotechnology
103, 273-280
Lee, J., 1997. Biological conversion of lignocellulosic
biomass to ethanol. Biotechnology 56, 1-24.
Lynd, L.R., Bothast, R.J., Wyman, D.E. 1991. Fuel
etanol from cellulosic biomass. Science 251:
1318-1323.
Martn, C., Galbe, M., Nilvebrant, N.O., Jnsson, L.J.,
2002. Comparison of the fermentability of
enzymatic hydrolyzed of sugarcane bagasse
pretreated by steam explosion using different
impregnating agents. Applied Biochemistry and
Biotechnol. 99, 699-716.
Maiorella, B. L. 1985 Industrial chemical, biochemical
and fuels: Ethanol. In: Comprehensive
biotechnology. The principles, applications and
regulations of biotechnology in industry,
agriculture and medicine. M.-M. Young (ed.),
Pergamon Press Ltd. University of Waterloo,
Ontario, Canada, Vol. 3. Chapter 43, 861-914.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-20 6/6
Makanjuola, D. B., Tymon, A. and Springham, D. G.
1992. Some effects of lactic acid bacteria on
laboratory-scale yeast fermentations. Enzyme
Microb. Technol. 14: 350-357.
Mc Millan, J.D., 1994. Pretreatment of lignocellulosic
biomass. In: Himmel, M.E., Baker, J.O.,
Overend, R.P (Eds), Enzymatic Conversion of
Biomass for Fuels Production. American
Chemical Society, Washington, DC, pp. 292-324
Palmqvist, E., Hahn-Hgerdal, B., Galbe, M., Larsson,
M., Stenberg, K., Szengyel, Zs., Tengborg, C.
and Zacchi, G. 1996a. Design and operation of a
bench-scale process development unit for the
production of ethanol from lignocellulosics.
Biores. Technol. 58: 171-179.
Pandey,A. Soccol,C.R. Nigam,P. And Soccol,V.T.
2000. Biotechnological potential of agro-
indistrial residues. Sugarcane bagasse.
Bioresour Technol. 74: 69-80.
Sun, Y., Cheng, J. 2002. Hydrolysis of lignocellulosic
materials for etanol production: review.
Bioresource Technology 83, 1-11.
Takagi,M., Abe, S., Suzuki, S., Emert, G.H., Yata,
N. 1977. A method for production of etanol
direcly from cellulose using cellulose and yeast.
Proceedings of Bioconversion symposium,
Delhi, 551-571.
Wyman, C.E., 1994. Etanol from ligcellulosic biomass:
Technology, economics, and opportunities.
Bioresource Technology 50, 3-6.
Zadrazil, F. Brunert, H., 1982. Solid state fermentation
of lignocellulose containing plant residues with
Sporotrichum pulverulentum Nov. and
Dichomitus squalens (Karst.) reid. European
Journal of Applied Mirobiology and
Biotechnology 16, 45-51
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-21 1/6
A General Application of a Direct Method for
Multivariable MPC Control Strategy in Chemical
Processes
A. Ahmad
1
and A. Wahid
2
1. Department of Chemical Engineering, Universiti Teknologi Malaysia, Johor
2. Department of Chemical Engineering, University of Indonesia, Depok.
Corresponding authors: arshad@fkkksa.utm.my and wahid@che.ui.edu
Abstract A general application of a direct method
for model predictive control strategy is proposed
for multivariable nonlinear control problem in
chemical processes. The aim is to provide a
solution to nonlinear control problem that is
favorable in terms of industrial implementation.
The scheme utilizes multiple linear models to cover
wider range of operating conditions. Depending on
the operating conditions, suitable models of 2x2,
3x3, and 4x4 processes are used in control
computations. Servo and regulatory controls of the
system are examined on constrained and
unconstrained conditions. Comparisons are made
to conventional controllers. The results confirmed
the potentials of the proposed strategy.
Keywords Tuning, model predictive control,
multivariable, nonlinear control, chemical process
I. INTRODUCTION
he cornerstone of MPC is the model [1]. It cause
MPC is called MBPC (model-based predictive
control). MPC uses models in 2 ways: using a reliable
model to predict effect of past control moves on P
future outputs, assuming no future moves, and using
the same model to compute the optimal M controller
moves. Implement first move and repeat procedure.
Predictive control is now one of the most widely
used advanced control methods in industry, especially
in the control of processes that are constrained,
multivariable and uncertain. A large number of
implementation algorithms, included industrial
predictive control applications [2] have a appeared in
the literature.
II. TUNING THE MPC CONTROLLER
Dynamic matrix control (DMC) [3] is the most
popular MPC algorithm used in the chemical process
industry today. Over the past decade, DMC has been
implemented on a wide range of process applications.
A major part of DMCs appeal in industry stems from
the use of a linear finite step response model of the
process and a simple quadratic performance objective
function. The objective function is minimized over a
prediction horizon to compute the optimal controller
output moves as a least-squares problem.
Tuning in conventional control strategy (P, PI, and
PID) is related to obtain an optimum setting of
controller parameters (K
c
, T
i
, and T
d
). Ziegler-Nichols,
Lopez, Ciancone, etc. [4] are some examples of
single-loop tuning in P, PI, and PID controllers.
Huang, et. al. [5] has proposed a direct method for
multi-loop PI/PID controller design based on
FOPDT/SOPDT model of each loop.
MPC controller has certain parameters setting to
achieve its optimum performance. During the time,
trial-and-error efforts have been done to reach this
goal until Shridhar & Cooper [6, 7] proposed a tuning
strategy for unconstraint SISO and multivariable
MPC. Dougherty 8and Cooper [8] proposed a non-
adaptive DMC tuning strategy (see Table 1) based on
all of FOPDT models in systems.
The principles of the multivariable DMC tuning
strategy [8] are:
1) T Sampling time. While a large T refers to a low
computation load, a small T refers to a properly
track the evolving dynamic behavior. Too slow of
a sampling rate will lead to information losses,
and too fast of a sampling rate could lead to
numerically sensitive procedures. Nevertheless,
this method allows values of T other than the
recommended value given in Table 1.
2) P (prediction horizon) and N (model horizon).
Both P and N have the same setting and are
related to the settling time of the slowest (the
largest time constant) sub-process in the
multivariable system. A large P improves the
nominal stability of the closed loop. Meanwhile,
a large N makes controller has long enough time
to avoid the instabilities that can otherwise result
since truncation of the model horizon
misrepresents the effect of controller output
moves in the predicted process variable profile.
3) M Control horizon. M equals to 63.2% of the
settling time of the slowest sub-process in the
multivariable system. This ensures M to be long
T
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-21 2/6
enough such that the results of the control actions
are clearly evidenced in the response of the
measured process variable.
4) - Controlled variable weights. The setting of
those parameter are free. However, in the most
case, they are set equal to one.
5) Move suppression coefficients. Its primary
role in DMC is to suppress aggressive controller
actions. When the control horizon is 1 (M = 1),
no move suppression coefficient is needed ( =
0). If the control horizon is greater than 1 (M >
1), then the analytical equation given in Table 1
is used.
Table 1 Non-adaptive DMC tuning strategy [8]
Approximate the process dynamics of all controller output to measured process variable pairs with FOPDT
models:
1. Select the sample time as close as possible to:
2. Compute the prediction horizon, P; and the model horizon, N:
3. Compute a control horizon, M:
M
4. Select the controlled variable weights, , to scale process variable units to be the same.
5. Compute the move suppression coefficients, :
6. Implement DMC using the traditional step response matrix of the actual process and the initial values of
the parameters computed in steps 16.
III. APPLICATION IN GENERAL PROCESSES
Dougherty and Cooper [8] have proved the non-
adaptive DMC tuning strategy in the three 2x2
processes (general transfer function, multi-tank, and
distillation column). Here, we shall illustrate
comparisons between PI/PID and MPC controller
strategy using Table 1 in the same and more complex
processes: 2x2 process, 3x3 process and 4x4 process
[5].
3.1 Wood and Berry distillation column (2x2 process)
Wood and Berry distillation column model [5] is
the most popular model of 2x2 process system. This
model consists of two matrixes: manipulated variable
matrix (2x2) and disturbance matrix (2x1).
Nevertheless, the first matrix (manipulated variable
matrix) is commonly used by researchers [5, 9].
Table 2 shows results of PID and MPC tuning in
Wood & Berry model. Responses of Wood & Berry
model using these settings are shown by Figure 1.
Although, MPC controller is not set in the optimum
setting, MPC controller performances are better than
PID controller performance. MPC controller tuning
has improved its performance by decreasing the
overshoot significantly. In unit step change test of y1,
the y2s overshoot sharply falls from 0.342 to 0.056,
or going down 84%.
Table 2 PID and DMC parameters setting in WB model (2x2 process)
Wood & Berry model
(2x2 process)
PID controller tuning
(Proposed 2 method)
DMC controller
tuning
) (
1 2 . 13
9 . 4
1 9 . 14
8 . 3
) (
) (
1 4 . 14
4 . 19
1 9 . 10
6 . 6
1 21
9 . 18
1 7 . 16
8 . 12
) (
) (
3
8
3 7
3
s F
s
e
s
e
s V
s R
s
e
s
e
s
e
s
e
s Y
s Y
s
s
s s
s s
B
D
+
+
+
+
+
+
=
Loop 1: ,
,
Loop 2: , ,
,
Ts = 1.5
P = 153
M = 35
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-21 3/6
0
0.5
1
1.5
y
1
0 10 20 30 40 50 60 70 80 90
-0.02
0
0.02
0.04
0.06
y
2
(a) (b) (c)
Figure 1 Response of multi-loop control for WB process (step change in y1): (a) PID controller performance using Proposed
2 tuning, (b) MPC controller performance using Ts=0.1, P=10, M=2 (default setting), (c) MPC controller performance using
Ts=1.5, P=153, M=35 (tuning setting)
3.2 OR (3x3) process
The Ogunnaike and Ray (OR 3x3) process [5] as
follows is considered (see Table 3). PID controller
tuning uses BLT-4 method instead of proposed
method by Huang et. al. [5]. In this case, responses of
OR (3x3) process are unstable using their proposed
method, though they are good using BLT-4.
As shown by Table 1, the tuning strategy uses
FOPDT (first-order plus dead-time) model to
calculate Ts, P, and M. Because of OR (3x3) model
has one SOPDT (second-order plus dead-time), this
transfer function has to be changed into FOPDT. By
using PRC (process reaction curve) from step change
testing of the SOPDT and applying a method that is
developed by Cecil L. Smith [4], the FOPDT is
obtained.
Figure 2 shows responses of multi-loop control for
OR (3x3) process: (a), (b), and (c) PID controller
performance using BLT-4 tuning (unit step change in
y1, y2, y3 respectively). They have very poor
performances. This disadvantage is improved by MPC
controller.
Next, we are going to compare between responses
using default setting and tuning setting. The results
are shown by Figure 3 and Table 4. Almost
performances by tuning setting are better than by the
default setting. Only two point where the default
setting is better than the tuning setting, they are
overshoot of y2 in y1s step change, and settling time
of y2 in y2s step change.
Table 3 PID and DMC parameters setting in OR model (3x3 process)
OR model (3x3 process) PID controller tuning
(BLT-4 method)
DMC controller
tuning
Loop 1: ,
,
Loop 2: ,
,
Loop 3: ,
,
Ts = 0.71
P = 91
M = 30
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-21 4/6
0 50 100 150
0
0.5
1
y
1
0 50 100 150
-1
0
1
y
2
0 50 100 150
-10
0
10
y
3
0 50 100 150
0
0.1
0.2
y
1
0 50 100 150
0
1
2
y
2
0 50 100 150
-10
0
10
y
3
0 50 100 150
-0.01
-0.005
0
y
1
0 50 100 150
-0.02
0
0.02
y
2
0 50 100 150
0
1
2
y
3
(a) (b) (c)
Figure 2 Responses of multi-loop control for OR (3x3) process: (a), (b), and (c) PID controller performance using BLT-4
tuning (unit step change in y1, y2, y3 respectively)
0
0.5
1
1.5
Response: Scenario1
Output: y1
Time (sec): 53.4
Amplitude: 1
y
1
-0.2
0
0.2
Response: Scenario1
Output: y2
Time (sec): 1.6
Amplitude: -0.154
y
2
0 10 20 30 40 50 60 70 80 90 100
-0.04
-0.02
0
0.02
Response: Scenario1
Output: y3
Time (sec): 2.02
Amplitude: -0.0387
y
3
Plant Outputs
0
0.5
1
1.5
Response: Scenario1
Output: y1
Time (sec): 47.5
Amplitude: 1
y
1
-0.2
0
0.2
Response: Scenario1
Output: y2
Time (sec): 2.84
Amplitude: -0.193
y
2
0 10 20 30 40 50 60 70 80 90
-0.04
-0.02
0
0.02
Response: Scenario1
Output: y3
Time (sec): 2.84
Amplitude: -0.0241 y
3
(a) (b)
Figure 3 Responses of multi-loop control for OR (3x3) process (step change in y1): (a) MPC controller performance using
Ts=0.1, P=10, M=2 (default setting); (b) MPC controller performance using Ts=0.71, P=91, M=30 (tuning setting)
Table 4 Performance of MPC controller of OR (3x3 process) model
Default Setting Tuning Setting
Step
Change
Controlled
Variable Settling time Overshoot Settling time Overshoot
y1 53.4 0.08 47.5 0.02
y2 -0.153 -0.193
y1
y3 -0.0386 -0.0241
y1 -0.156 -0.138
y2 21.7 0.01 31 0.01
y2
y3 -0.151 -0.0552
y1 -0.0303 -0.0273
y2 -0.0852 -0.0614
y3
y3 6.52 0 3.18 0
3.3 Alatiqi case 1 (4x4) process
Consider alatiqi case 1 (A1 4x4) process (see
column 1 in Table 5). Alatiqi process has 16 empirical
models that consist of three FOPDT models, two
models having zeros, and eleven SOPDT models. All
of empirical models have to be changed in FOPDT
model to calculate MPC controller tuning. We use
some PRCs (process reaction curves) from step
change testing of the SOPDTs and the models having
zeros, and apply a method that is developed by Cecil
L. Smith [3] btain the FOPDT.
PI controller shows the poor performance (see
Figure 4). Also, the controller performance as shown
by Figure 5 is extremely poor, because the responses
are unstable. So, in this case, the default setting can
not be used in MPC controller. Inevitable, we have to
use the tuning setting in this case.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-21 5/6
Table 5 PID and DMC parameters setting in Alatiqi case 1 model (A1 4x4 process)
Alatiqi case 1 model (A1 4x4 process)
PI controller tuning
(Lee et al. method)
DMC controller
tuning
Loop 1: ,
Loop 2: ,
Loop 3: ,
Loop 4: ,
Ts = 1.62
P = 191
M = 52
0 20 40 60 80 100 120 140 160 180 200
0
0.5
1
y
1
0 20 40 60 80 100 120 140 160 180 200
-0.01
0
0.01
y
2
0 20 40 60 80 100 120 140 160 180 200
-0.5
0
0.5
y
3
0 20 40 60 80 100 120 140 160 180 200
-2
-1
0
y
4
0
1
2
y
1
-0.5
0
0.5
y
2
-0.05
0
0.05
y
3
0 10 20 30 40 50 60 70 80 90
-0.5
0
0.5
y
4
Plant Outputs
(a) (b)
Figure 4 Responses of multi-loop control for alatiqi case 1 (A1 4x4) process: (a) PI controller performance using Lee et. al.
tuning (unit step change in y1); (b) MPC controller performance using tuning setting Ts=1.62, P=191, M=52 (unit step
change in y1)
-5
0
5
x 10
5
y
1
-2
0
2
x 10
4
y
2
-100
0
100
y
3
0 10 20 30 40 50 60 70 80 90 100
-4
-2
0
x 10
4
y
4
Plant Outputs
Figure 5 Alatiqis Responses of MPC controller using
default setting
IV. CONCLUSSIONS
MPC controller offers better control performances
than PI/PID controller, especially in multivariable
processes. Application of MPC controller in the three
more complex models produces the fantastic
performance. To achieve an optimum performance of
MPC controller, non-adaptive DMC controller tuning
can be used. This method has proved that the tuning
setting has the optimum performance. In the complex
processes like WB model (2x2 process) and OR
model (3x3 process), the default setting can be used in
MPC controller. But, in more complex process likes
Alatiqi case 1 model (A1 4x4 process), the default
setting produces very poor performance. So, we must
use the tuning setting.
REFERENCES
[1] D.W. Clarke. Adaptive Predictive Control. A
Rev. Control. Vol. 20, pp. 83-94, 1996
[2] S.J. Qin and T.A. Badgwell, (2003). A survey of
industrial model predictive control technology.
Control Engineering Practice. (11): 733 764
[3] C. R. Cutler & D. L. Ramaker. Dynamic matrix
controla computer control algorithm.
Proceedings of the JACC 1980. San Francisco,
1980.
[4] T. Marlin. Process Control: Designing Processes
and Control Systems for Dynamic Performance.
2nd Edition,McGraw-Hill, New York, 2000.
[5] Huang et. al. A direct method for multi-loop
PI/PID controller design. Journal of Process
Control (13): 769786, 2003.
[6] R. Shridhar & D. J. Cooper. A tuning strategy for
unconstrained SISO model predictive control.
Industrial & Engineering Chemical Research.
(36): 729746, 1997.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-21 6/6
[7] R. Shridhar & D. J. Cooper. A tuning strategy for
unconstrained multivariable model predictive
control. Industrial & Engineering Chemical
Research. (37): 40034016, 1998.
[8] Danielle Dougherty and Doug Cooper. A
Practical Multiple Model Adaptive Strategy for
Multivariable Model Predictive Control. Control
Engineering Practice. (11): 649 664, 2003.
[9] B. Huang., Steven X. Ding, and Nina Thornhill.
Alternative solutions to multi-variate control
performance assessment problems. Journal of
Process Control (16): 457471, 2006.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-22 1/4
Biogrease Using Modified Palm Oil as Base Oil and
Thickener Lithium Soap
Sukirno *, M.Nasikin, B.Heru, Rizqon F, Marius, Dizzi.
* Fac. of Engineering, University of Indonesia, Jl. Salemba Raya 4, Jakarta 10430
Tel. 7863516, fax. 7863515 email : sukirnos@che.ui.edu
Abstract Biogreases without additives, have been
prepared using biolubricant, as the base oil. This
bio lubricant is a modified palm oil which can be
synthesized from palm oil via transesterification and
epoxidation reactions. The thickening agent used in
this grease making study was lithium 12-
hydroxystearate. The making procces of the
biogrease was carried out in a closed flanged flask
or autoclave. Characterizations of the biogrease
were carried out using Penetration Test, Dropping
Test. It was found that lithium biogrease NGLI 2
can be prepared using thickener less that was
needed by mineral oil grease. Performance test by a
Gear Load-carrying test showed that the biogrease
products has better antiwear performance than that
of lithium grease using mineral oil HVI 160 S as the
base oil.
Keywords Biogrease, lithium soap, base oil.
Penetration Test, Dropping Point Test, Gear Load-
carrying Test.
I. INTRODUCTION
ubrication is proccess of interposing a lubricating
material between two surfaces in relative motions in
order to reduce friction
[1]
. Liquid lubricants are the most
known type of lubricanting material, however semi
solid lubricant or grease is widely used in many
lubrication systems due to enomical consideration.
Lubricating greases are, in general, produced by
dispersing a solid thickening agent in a lubricating
fluid. The base oil selected in formulating a grease
should have the same characteristics, as if the
equipment is to be lubricated by the fluid lubricant.
Hence a fundamental difference between a formulated
fluid lubricant and a similarly fully formulated grease,
is the presence in of the thickening agent. These
thickeners are usually soaps, such as lithium, sodium
and calcium salts of long chain fatty acids, and the
more recently introduced thickeners, such as clays,
polyureas
[3]
. The same as lubricating oil, greases also
contain some additives in order to enhance its
performance and to protect the grease and lubricated
surfaces.
[1]
. The most common additives found in
grease are antioxidants to prolong the life of a grease,
anti-corrosion agents to protect metal against attack
from water, sulphides or corrosive elements, extreme
pressure agent to guard against scoring and galling,
antiwear agent to prevent abrasion and metal to metal
contact etc. So base oil, thickener dan addives are
blended to make a grease Typical grease composition
are shown in table 1.
Table l Typical grease composition
(%)
Base oil 75-95
Thickener 5-20
additives 0-10
Actually, the component that perform as lubricant
is the base oil
[3]
. The thickener gives grease its
characteristic rigidity or consistency and is sometimes
thought of as a three-dimensional fibrous network or
sponge that holds the oil in place. During the grease
in use, when the lubricant film between wearing
surfaces thins, the friction heat softens the adjacent
thickener, which expands and releases oil to restore
film thickness. Soap thickeners not only provide
consistency to grease, they affect desired properties
such as water and heat resistance and pumpability.
Fluid lubricants used to formulate grease are
normally petroleum or synthetic oils. Synthetic oils are
higher in cost, so that the majority of greases on the
market are composed of mineral oil. With growing
environmental concerns, vegetable oils are also being
used in applications requiring nontoxic or
biodegradable greases. In America, they are developing
soybean oil. For Indonesia, as the main producer of
palm oil, we should develop palm oil.
In our previous research, we had successfully
prepared a biolubricant via transesterification and
epoxidation reaction of palm oil to a lubricant which
has better oxidation stability than its feed, a RBDPO
olein. This modified palm oil is called called EPOME
Gliserol
[2]
which is biodegradable, and contain no
aromatic, nitrogen, sulfur, and is expected to be used in
foodgrade lubricant formulation. In this present reseach,
the biolubricant is used as base oil for grease making.
The biogrease product is supposed to be better in its
antiwear propertiy, since its has many active
functionality to protect friction surfaces.
The most important feature of a grease is its
rigidity or consistency. A greases consistency is its
resistance to deformation by an applied force. A grease
that is too stiff may not feed into areas requiring
lubrication, while a grease that is too fluid may leak
out. The measure of rigidity or consistency is called
penetration which is measured by penetration test. To
measure penetration, a cone of given weight is allowed
L
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-22 2/4
to sink into a grease for 5 seconds at a standard
temperature of 25 C). The depth, in tenths of a
millimeter, to which the cone sinks into the grease is
the penetration. A penetration of 100 would represent a
solid grease while one of 450 would be semifluid The
NLGI has established consistency numbers or grade
numbers, ranging from 000 to 6, corresponding to
specified ranges of penetration numbers. Table 1 lists
the NLGI grease classifications along with a description
the consistency of each classification.
Tabel 2 NLGI Number
NLGI Worked Penetration (Grade)
at 25
0
C (0,1mm)
Virtual Assessment of
Deformability
000 445-475 Very soft
00 400-430 Very soft
0 355-385 Soft
1 310-340 Creamy
2 265-295 Almost solid
3 220-250 Hard
4 175-205 Hard
5 130-160 Very hard, as soap
6 85-115 Very hard, as soap
The purpose of this research is to prepare
biogrease of NGLI 2 which is normally used for gear
and wheel bearing application. In this preliminary study
of making grease, the amount of the thickener is varied
to observe their rigidity or consistency and dropping
points in order to make a graph shown the effect of the
amount of thickener to the biogrease characteristics,
and its lubrication performance in comparison with
similar grease using mineral oil as base oil.
II. EXPERIMENTAL
Materials used
(a) Base oil
The base oil used was a bio lubricant synthesized
from palm oil. It was the product of previous research
and its properties are given in Table 1. The Mineral oil
HVI 160 Pertamina product was also used as base oil
comparison.
Table 1 Properties of base oils
Appearance Yellow pale Yellow
Specific gravity [-] ASTMD-1289 0.91 0.85 0.8
Viscosity @40C cSt ASTMD-445 35 38.9 96
Viscosity @100C cSt ASTMD-446 6.9 8-9 11
Viscosity Index [-] ASTMD-2270 >100 >100 100
Pour point C ASTMD-97 0 8-15 -9
Oxid. Test ( Vis) Bulk Oxidation 5.6 21.5 7.8
Test Method
Biolubricant
(Modified
RBDPO)
RBDPO
M.O
HVI 160
S
Characterization
(b) Thickener
Thickener lithium soap was obtained from the
reaction of fatty acid 12-hydroxystearate (melting point
77.5C).and lithium hydroxide.
Preparation of greases
A mixture of lithium 12-hydroxystearate and
approximately 85% of the bio lubricant to be used in
the manufacture of the grease was heated in a flanged
flask or autoclave, equipped with a stirrer, pressure
indicator, oil heater thermometer, to 120C and stirred,
as shown in figure 1. The lithium hydroxide was added
slowly until solution of the soap occurred (170C).
Heating was continued until the soap melted (200C).
Next, the solution was cooled in a cooler to
approximately 70C with vigorous stirring. atmosphere.
The remainder of the base oil, is then added to give the
final product. Stirring was continued for a further about
1 h, and the product, now a creamy grease, was milled
in Miller..
In this experiment, the amount of the thickener is
varied, and additives has not been added.
Fig.1 Autoclave
Biogrease Product Characterizations
Samples of the greases were tested with
Penetrometer (ASTM D 217), Dropping Point (ASTM
D 566) Water washout (ASTM D 1264).
Biogrease Product Performance Test
Performance tests of biogreases were conducted
using Gear Load-carrying tests. The method used to
assess the load-carrying characteristics of the greases in
this experiment was the gear load-carring test. In this
test, a pair of gear were used. Each test was carried out
at ambient temperature with 10 hours runs at 25 rev/s
and the applied load of 10 kg. The wear debris is
measure by AAS.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-22 3/4
Fig.2 Gear load-carrying test
III. RESULTS AND DISCUSSION
Product Appearance
Figure 3 shows a grease product prepared in
autoclave, as seen under ordinary camera. It is creamy
and the texture is soft and sticky.
Fig.3 Biogrease
Biogrease rigidity
Fig.4 shows the plot between penetration and
composition of the thickener lithium soap obtained
from the biogrease of this experiment. In the figure,
there is also given a plot of a grease using mineral oil,
as a comparison. From the figure can be seen that,
grease consistency increase as the amount of thickener
increase, and also with the same amount of thickener
the biogrease is more rigid than that of petroleum
grease. This means that to produce the same level of
rigidity, the biogrease need smaller amount of the
thickener than that of petroleum grease.
100
150
200
250
300
0 5 10 15 20 25
%Lithium Soap
P
e
n
e
t
r
a
t
i
o
n
(
0
,
1
m
m
)
Lithium Grease HVI 160 Lithium Biogrease
Fig.4 Penetration Value of Li Biogrease
The Biogrease Dropping Point.
Dropping point is an indicator of the heat
resistance of grease. As grease temperature rises, the
grease gets softer, until liquefies and the rigidity is lost.
Dropping point is the temperature at which grease
becomes fluid enough to drip. The dropping point
indicates the upper temperature limit at which grease
retains its structure, not the maximum temperature at
which a grease may be used.
The dropping point measurement of the biogrease
are shown in fig 5. Dropping point increases as the
amount of thickener increases, and for the same amount
of thickener lithium soap, the dropping point of
biogrease is higher than that of mineral oil grease.
100
120
140
160
180
200
220
240
0 5 10 15 20 25
%Lithium Soap
D
r
o
p
p
i
n
g
p
o
i
n
t
C
Li Biogrease Li Grease HVI 160
Fig.5 Dropping Point of Li Biogrease
This experiment also tried to make semibio grease,
where the base oils are mixture of biolubricant and
mineral oil. Figure 3 shows the result of the dropping
point test of semibio grease with various base oil
compositions. The dropping point of semiobio grease
increases as the % of biolubricant increases.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-22 4/4
Li Grease semibio (Biolubricant +HVI 160S)
0
50
100
150
200
250
0 20 40 60 80 100
% Bio lubricant (Modified Palm Oil)
D
r
o
p
p
i
n
g
p
o
i
n
t
(
d
e
r
C
e
l
s
i
u
s
)
Fig.6 Dropping Point of Lithium Semibio Grease
Table 6 lists the results of quality tests of biogrease
in comparison with that of the mineral oil grease.
Lithium biogrease is better than lithium petroleum
grease, in case of their dropping point. But it is inferior
in water washout tests. This test is the ability of grease
to withstand the effects of water with no change in its
ability to lubricate. Water may suspend the oil in the
grease, forming an emulsion that can wash away or, to a
lesser extent, reduce lubricity by diluting and changing
grease consistency and texture. Rusting becomes a
concern if water is allowed to contact iron or steel
components.
Table 2 Characteristic Comparison of BioGrease and
Mineral Oil Grease
Li Soap (%) 15 15 10
Penetration (0,1 mm) 220.7 148 229
NLGI number 3 5 3
Dropping point (
0
C) 221 230 210
Water washout (%) 5 <10* <15*
Li Grease
( HVI 160 S
Pertamina)
Lithium Biogrease
Gear Load-carrying test
The result of performance tests of the biogrease
using gear load-carrying test are shown in Figure 7. In
this figure the amount of wear debris (ppm of Fe) were
plotted against % biolubricant. It can be seen that, as
the composition (%) of biolubricant increases the wear
amount (in ppm of Fe) decreases. It means that
biolubricant can improve the performance of petroleum
grease when is used in grease formulation.
0
50
100
150
200
250
Lithium 15% Lithium 15% Lithium 15% Lithium 15% Lithium 10%
Biolubricant
40% HVI
Biolubricant
60% HVI
Biolubricant
80% HVI
Biolubricant
100% HVI
Biolubricant
100% HVI
p
p
m
F
e
Fig.7 Wear Rate of Li Biogrease using Gear load-
carrying test.
IV. CONCLUSIONS
From this preliminary study of making biogrease
with modified palm oil biolubricant as the base oil,
several conclusions can be summarized as follow:
a. The amount of lithium soap needed is .10 % of
total weight to produce a biogrease NGL 2 using
modified palm oil as the base oil. To reach the
same level of rigidity consistency, the biogrease
without additives needs smaller amount of the
thickener than that of petroleum grease.
b. In gear load-carrying tests, the biogrease without
additives showed better antiwear performance than
that of petroleum grease, with HVI 160 S as the
base oil.
ACKNOWLEDMENTS
This work was funded by RUSNAS Industri Hilir
Kelapa Sawit. The Authors would like to express their
appreciation all who supports to this study.
REFERENCES
1. OBrien, J.A. , Lubricating Oil Additives, Booser, E.R.,
editor., CRC Handook of Lubricating (Theory and
Practice Tribology). Volume 2, Theory and Design, CRC
Press Inc. Boca Raton, Florida, 1983.
2. Fenjerry, Y. "Preparation and Characterization of Epome
Gliserol dan Epome Monoalcohol as Foodgrade
Lubricant", Bachelor Thesis, Chemical Engineering
Department, University of Indonesia, Depok, 2005.
3. H. B. SILVER and I. R. STANLEY, The Effect of The
Thickener on the Efficiency of Ioad-Carryng Additives in
Greases, TRIBOLOGY international June, pp.113-118,
1974.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-23 1/5
Model for Oxidation and Combustion of Ethanol in Wide
Range of Pressures, Temperatures and Equivalence Ratios
Yuswan Muharam and Nonni Soraya Sambudi
Department of Chemical Engineering, Faculty of Engineering, University of Indonesia,
Kampus UI Depok 16424
Phone 021-7863516, fax. 021-7863515, email: muharam@che.ui.edu
Abstract Ethanol as an alternative fuel can
be used either in a mixture or single component. The
advantage of ethanol over gasoline as fuel is the
minimum standard pollutant emitted from its
burning, such as carbonmonoxide. Ethanol oxidation
and combustion processes use 372 elementary
reactions and yield various intermediates and
products.
This computational study has several purposes to
find out, such as ignition delay times, emitted
pollutants and influences of pressure, temperature
and equivalence ratio on ethanol oxidation and
combustion. Hence, a comprehensive and
representative chemical kinetic model for ethanol
oxidation and combustion is needed to fulfill the
purposes of this study. This model is expected for
high validation and represents the real condition of
ethanol oxidation and combustion. The
computational results are compared to the
experimental data from flow reactors and jet stirred
reactors for concentration profiles, and shock tube
for ignition delay times. Good agreement was found
for the validation with experimental data.
Mechanism analyses were doing by applying the
sensitivity and reaction flow analyses. Sensitivity
analysis is used to identify rate-limiting reaction
steps. Reaction flow analysis calculates the
percentage of reaction contribution to the formation
or consumption of chemical species. Sensitivity
analysis shows that OH O+ H O
2
+ is the most
sensitive reaction for ignition delay times.
Ethanol oxidation and combustion were
simulated within the temperature range of 1000-2000
K, the pressure range of 1-15 atm and the
equivalence ratio ( ) range of 0.5-2.0. The simulation
was performed using Chemkin 3.7.1 software.
Keywords ethanol oxidation and combustion, kinetic
model, ignition delay times
I. INTRODUCTION
he usage of fossil fuels suffers pollution problems
that affect the environment. This has become the
main issue of energy resource exploitation. Energy
alternative rises as the solution for less standard
pollutant and sustainable energy. Bioethanol is an
energy alternative which usage is expected to solve the
energy problems.
Based on the National Energy Mix, the Government
has planned the usage of biodiesel and bioethanol in
amount of approximately 2% of overall national fuels
usage in 2010, and will rise up to 5% in 2025 [1].
Previous chemical kinetic modeling studies of
ethanol oxidation were performed by Natarajan and
Bhaskaran [2]. The following modeling works focused
on problems of ethanol ignition delay in shock tubes,
and ethanol laminar flame speeds in burners and jet
stirred reactors [2,3,4,5]. The time needed from the start
of injection and the start of combustion is referred to as
the ignition delay.
Egolfopoulos [4] modeling study demonstrated a
good agreement with Natarajan [2] ignition delay data
and found (
M) OH( H C
5 2
+ M) OH( CH CH
2 3
+ +
) as the
most sensitive reaction for ignition delay. Curran [6] did
not find it in his modeling study at 1100-1900 K and 2-
4.5 bar. Detailed mechanism is generally validated with
experimental data from certain range of operating
condition [7].
So far, the modeling of ethanol oxidation and
combustion had not achieved wide validation range.
Hence, a comprehensive mechanism of ethanol
oxidation and combustion in wide range of operating
condition (pressure, temperature and equivalence ratio)
is needed to be validated with various experimental data
and represents the actual combustion system.
II. BASIC THEORY
Chemical reactions occur when molecules of one
species collide with molecules of another species and,
for some of these collisions, one or more new molecules
T
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-23 2/5
will be created. In the chemical reaction the atoms of
reacting molecules are redistributed in the new
molecules. To achieve this, the reacting molecules must
have sufficient kinetic energy so that their chemical
bonds can be broken during the impact and other bonds
can be formed. As the energy of these bonds depends on
the nature of the atoms and on geometrical factors, the
energy content of the products of the collision may be
different from the energy content of the colliding
molecules. This is the basis for heat being released or
absorbed in chemical reactions [8].
II.1 Temperature Dependence of Rate Coefficients
The reaction rate coefficient is given by the
Arrhenius law:
=
RT
E
A k
a
exp'
(1)
where E
a
is the activation energy. Not all molecular
collisions will result in reaction, but only those with
kinetic energy higher than the energy needed to break
the bonds of the reactant molecules. This energy barrier
is the activation energy. Its maximum value corresponds
to the bond energies in the molecule. In dissociation
reactions, the activation energy is approximately equal
to the bond energy being broken. The value of
activation energy may also be much smaller or even
zero [9].
Since most elementary binary reactions exhibit
Arrhenius behavior over modest ranges of temperature,
the temperature dependence can usually be incorporated
with sufficient accuracy into the exponential alone.
However, for the large temperature ranges found in
combustion, non-Arrhenius behavior of the rate
coefficient tends to occur, particularly for processes that
have a small energy barrier. Therefore, it is necessary to
adopt a modified Arrhenius form which expresses the
impact of temperature on the rate coefficient [10],
=
RT
E
AT k
b a
exp
. (2)
II.2 Pressure Dependence of Rate Coefficients
The Lindemann mechanism illustrates the fact that
reaction orders of complex (nonelementary) reactions
depend on the conditions chosen. However, the
Lindemann mechanism itself is a simplified model.
If the rate law of a unimolecular reaction is written
as
[A] d[P]/d
a
k t =
, then the rate coefficient k depends on
pressure and temperature. The theory of unimolecular
reactions yields fall-off curves, which describe the
pressure dependence of k for different temperatures. The
logarithm of the rate coefficient is usually plotted versus
the logarithm of the pressure [9].
Typical fall-off curves are shown in Fig. 1. At very
high pressures
) [M] [M]/(
u a - a u
k k k k k + =
tends to the
limit
k
, i.e., the rate coefficient becomes independent
of the pressure. At very low pressure, the rate
coefficient k is proportional to [M] = p/RT, resulting in a
linear dependence. Similarly, if the effective activation
energy is low, the reaction rate coefficient k will
decrease with temperature [9].
Fig.1. Fall of curves for unimolecular reaction
3 3 6 2
H C H C H C
& &
+
[9]
II.3 Ignition Delay Time
Ignition delay is characteristic for radical chain
explosions. During the ignition delay period, the radical
pool population is increasing at an exponential rate. Yet,
the amount of fuel consumed, and hence the amount of
energy liberated, is too small to be detected. Thus,
important chemical reactions (chain branching,
formation of radicals) take place during the induction
time, whereas the temperature remains nearly constant
[9].
Finally, the radical pool becomes large enough to
consume a significant fraction of the fuel, and rapid
ignition will take place. The precise definition of
induction time depends on the criterion used
(consumption of fuel, formation of CO, formation of
OH, increase of pressure in a constant volume vessel,
etc.) [9].
Due to the temperature dependence of the
underlying elementary reactions, the ignition delay time
depends strongly in the temperature. This is shown in
Fig. 2 for several hydrocarbon-air mixtures [9].
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-23 3/5
Fig.2. Calculated (line) and measured (points) ignition delay
times in hydrocarbon-air mixtures [9]
III. EXPERIMENTAL RESULTS
III.1 Mechanism Validation
A mechanism consisting 372 elementary reactions
for ethanol oxidation and combustion has been
developed in this study. The mechanism validation was
performed by comparing calculation results to
experimental measurements for shock tubes, jet stirred
reactors and flow reactors available in literature.
10
100
1000
6 6,5 7 7,5 8
10000/T (K
-1
)
I
g
n
i
t
i
o
n
d
e
l
a
y
t
i
m
e
s
(
m
i
k
r
o
d
e
t
i
k
)
Fig. 3. Comparison between shock tube ignition delay data as
investigated by Natarajan and Bhaskaran [2] and the
numerical calculations using the chemical kinetic model.
Experimental conditions: 2.5% C2H5OH, 7.5% O2, and 90%
Ar, = 1.0,
P = 1 0.2 and 2.0 0.2 atm
As shown in Fig. 3, ignition delay time is shorter at
higher pressure and temperature. This behavior was
reproduced very well by the model.
Figure 4 shows temperature behavior for ignition
delay time in shock tube. Pressurized wave resulted in
shock tubes will increase the system pressure hence
temperature. The rise of pressure and temperature will
improve system reactivity and radical pool. Formation
of large enough radical pool will consume significant
fraction of the fuel and rapid ignition will take place.
Fig.4. Temperature behaviour from ignition delay time in
shock tube
Aboussi [11] investigated ethanol oxidation in jet
stirred reactors. The experimental data were used to
validate the kinetic model by comparing it with
numerical calculations.
0,000001
0,00001
0,0001
0,001
0,01
0,1
0 0,05 0,1 0,15 0,2 0,25 0,3
Residence time (s)
F
r
a
k
s
i
m
o
l
C2H5OH sim C2H4 sim CO sim CO2 sim
C2H5OH exp C2H4 exp CO exp CO2 exp
Fig.5. Ethanol-O
2
-N
2
(0.3%0.9%98.8%) oxidation
in a jet stirred reactor at = 1.0, 1.0 atmosphere,
and a nominal temperature of 1056 K. Measurements and
predictions for C
2
H
5
OH, C
2
H
4
, CO and CO
2
Figure 5 shows the comparison between numerical
calculations for C
2
H
5
OH, C
2
H
4
, CO and CO
2
. Overall,
the modeling results show good agreement with the
species concentration as the residence time was varied.
Norton and Dryer [12] investigated ethanol
oxidation in turbulent flow reactors. The experimental
data were used to validate the kinetic model by
comparing it with numerical calculations using the
chemical kinetic model.
0
0,02
0,04
0,06
0,08
0,1
0,12
0 0,02 0,04 0,06 0,08 0,1 0,12
Residence time (s)
F
r
a
k
s
i
m
o
l
C2H5OHexp O2 exp COexp H2Oexp H2 exp CO2 exp
C2H5OHsim O2 sim COsim H2Osim H2 sim CO2 sim
O
2
0,8
C
2
H
5
OH 2
H
2
O
CO
H
2
CO
2
Fig.6. Comparison between experimental flow reactor
oxidation data for = 1.24 as investigated by Norton and
Dryer [12] and the numerical calculations. The numerical
results were time shifted by 40 msec. Experimental
conditions: 5.81% C2H5OH, 14.07% O
2
, and 80.12% N
2
,
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-23 4/5
Reynolds Number = 4900, P = 1atm, and T
in
= 1100 K.
Numerical simulations shown for C
2
H
5
OH, O
2
, CO, CO
2
, H
2
,
and H
2
O.
Fig.6. shows good agreement between experimental
data and numerical results for C
2
H
5
OH, O
2
, CO, CO
2
,
H
2
and H
2
O in temperature of 1100 K and equivalence
ratio of 1.24.
III.2 Mechanism Analyses
III.2.1 Sensitivity Analysis
A sensitivity analysis is used to identify rate-limiting
reaction steps. Numerical calculations for sensitivity
analysis were carried out by changing the rate
coefficient of a particular reaction and calculating the
OH concentration.
-0,6 -0,4 -0,2 0,0 0,2 0,4 0,6 0,8 1,0 1,2
1
= 0,5
= 2,0
Fig.7. Sensitivity coefficients of the important reactions for
ignition delay time in lean and rich fuel mixture, based on
Dunphy dan Simmie[3] experimental condition in shock tube.
P = 3.4 bar, T = 1430 K,
1
= 0.5 (1.25% C
2
H
5
OH, 7.5% O
2
and 91.25% Ar),
2
= 2.0 (2.5% C
2
H
5
OH, 3.75% O
2
and
93.25% Ar)
Figure 7 shows sensitivity coefficients of the
important reactions for ignition delay time. Six most
sensitive reactions for lean and rich fuel mixture are:
(+) OH O+ H O
2
+
(+)
2 3
HO CH + OH O CH
3
+
(+) M HCO+ M CO H + +
(+)
2
O HCO+
2
HO CO+
(+)
M O CH CH
2 3
+ M H HCO CH
3
+ +
(-)
M O CH CH
2 3
+ M O CH CH
2 3
+ +
A sensitivity coefficient might be positive or
negative. A positive sensitivity coefficient indicates a
higher OH concentration and an increased overall
reaction rate, and a negative sensitivity coefficient
indicates a lower OH concentration and a decreased
overall reaction rate of the system.
-0,5 0 0,5 1 1,5
1
CH3+HO2=CH3O+OH
O+OH=O2+H
H+HO2=H2+O2
CH3+CH3(+M)=C2H6(+M)
CH3+HO2=CH4+O2
C2H5OH(+M)=CH3+CH2OH(+M)
C2H5OH(+M)=C2H4+H2O(+M)
C2H5OH+OH=C2H4OH+H2O
C2H5OH+CH3=CH3CHOH+CH4
C2H5OH+HO2=CH3CHOH+H2O2
CH3CH2O+M=CH3HCO+H+M
CH3CH2O+M=CH3+CH2O+M
Fig.8. Sensitivity coefficients of the important reactions for
ethanol oxidation in a flow reactor. Experimental condition:
=0,61 (5.65% C
2
H
5
OH, 27.86% O
2
and 66.49% N
2
), P = 1
atm and T
in
= 1092 K
Figure 8 shows sensitivity analysis results for =
0.61 in flow reactors. Several most sensitive reactions
ethanol oxidation in flow reactor are:
(+) OH O+ H O
2
+
(+)
2 3
HO CH + OH O CH
3
+
(-)
2 3
HO CH +
2 4
O CH +
(-)
2
HO H+
2 2
O H +
(-)
M O CH CH
2 3
+ M O CH CH
2 3
+ +
(+)
M O CH CH
2 3
+ M H HCO CH
3
+ +
III.2.2 Reaction Flow Analysis
A reaction flow analysis calculates the percentage of
reaction contribution to the formation or consumption of
chemical species.
+O2
5,2%
+O2, M
83,7 %
+OH, HO2
97,8 %
+H, OH, O
64,1 %
+OH, O, H
68,3%
+OH
29,1%
+O
24%
+O2
11,1%
+O, O2
40,3%
+OH, H, O
65,1%
+O
27,8%
+M
46,8%
+O2
46%
+O
64,6%
+H
11,2 %
CO
HCO
CH2O
CH2OH CH3
CH3OH
CH2CO
CH4
C2H4
CH2HCO
CH3HCO
CH3CH2O CH3CHOH
CO2
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-23 5/5
Fig.9. The pathways of CO
2
formation in flow reactor at P = 1
atm, equivalence ratio () = 0.61 and T = 1092 K
Figure 9 shows carbondioxide formation from its
reaction with OH and HO
2
radical.
OH CO+ H CO
2
+
92.6%
2
HO CO + OH CO
2
+ 5.2%
Carbonmonoxide is resulted from radical formyl
oxidation, formyl decomposition with third body, and
oxidation and decomposition of CH
2
HCO.
2
O HCO +
2
HO CO+ 75.7%
M HCO+ M CO H + + 8%
2 2
O HCO CH + OH CO O CH
2
+ +
3.3%
HCO CH
2
CO CH
3
+
1.9%
IV. CONCLUSIONS
Validation of the mechanism for oxidation and
combustion of ethanol in shock tubes and flow and jet-
stirred reactors shows good agreement. The elementary
reaction OH O+ H O
2
+ is shown as the most
sensitive reaction for ignition delay time at 1430 K, 3.4
and 5 bar, and the equivalence ratios of 0.5 and 2.0. It is
also the most sensitive reaction for ethanol oxidation
rate in flow reactor at 1092 K, 1 atm, and the
equivalence ratio of 0.61.
Ignition delay time in shock tube is shorter at high
pressure and fuel lean condition. Lowest species
concentrations in flow and jet sstirred reactor are found
at low equivalence ratio, low pressure and temperature.
REFERENCES
[1] Prasetyo E. Energi Alternatif Ramah Lingkungan.
Accesed on 22 Mei 2006, from Pikiran Rakyat.
http://www.pikiranrakyat.com/cetak/2005/0705/15/otokir/ut
ama01.htm, 2005.
[2] Marinov N.M . A Detailed Chemical Kinetic Model for
High Temperature Ethanol Oxidation. International
Journal of Chemical Kinetic, pp.183, 1999.
[3] Dunphy M.P, J.M. Simmie. Journal Chemical Society
Faraday Trans 87 pp.16911695, 25492559, 1991.
[4] Egolfopolous F.N, D.X. Du, C.K. Law. Twenty-Fourth
Symposium (International) on Combustion. The
Combustion Institute, pp. 833, 1992.
[5] Dagaut P, M. Cathonnet, J.C. Boettner. Journal Chemistry
Physic 89, pp.867884, 1992.
[6] Curran H.J, et al. Twenty-Fourth Symposium
(International) on Combustion. The Combustion Institute,
pp. 769, 1992.
[7] Held T.J, F.L. Dryer. A Comprehensive Mechanism for
Methanol Oxidation. International Journal Chemical
Kinetic 30, pp.805, 1998.
[8] Muharam Y. Detailed Kinetic Modelling of the Oxidation
and Combustion of Large Hydrocarbons Using an
Automatic Generation of Mechanisms. Dissertation,
University of Heidelberg, Heidelberg, pp.1-27, 2005.
[9] Warnatz J, U. Maas, R.W. Dibble. Combustion, Physical
and Chemical Fundamentals, Modelling and Simulation,
Experiments, Pollutant Formation 3rd ed. Heidelberg:
Springer, 2001.
[10] Glassman I. Combustion 3rd edition. New York:
Academic Press, 1996.
[11] Aboussi B. Dissertation, Orleans, 1991.
[12] Norton T.S, F.L. Dryer. An Experimental and Modeling
Study Of Ethanol Oxidation Kinetics in an Atmospheric
Pressure Flow Reactor. International Journal Chemical
Kinetic 24, pp.319344, 1992.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-24 1/3
WATER AND AIR CIRCULATION SYSTEM IN
TRIMARAN FISHING VESSEL LIVE FISH HOLDS
Sunaryo, Anita Muslih
Naval Architecture Study Program, Mechanical Engineering Department, Faculty of Engineering, University of
Indonesia.
Kampus UI, Depok 16424, tel: 021-7270032, fax: 021- 7270033
naryo@eng.ui.ac.id
ABSTRACT- Indonesia has abundant of natural
resources including fish and aquaculture, but so
far the effort for exploiting it has not optimal yet,
such as in providing sufficient fishing fleet for the
national fishing industry, and the enabling of
traditional fishermen community to improve their
living standard. For this purpose sea fish farming
is introduced to some potential fishermen
communities. One type of the fish being farmed is
grouper, which has high market value to be
exported to the Far-East. But the fish in demand
in the market is live fish, therefore triggered by
this challenge as a continuation of the previous
research, this research is carried out focusing in
the arrangement of the water and air circulation
system in the Trimaran fishing vessel live fish hold,
so that the fish being transported could be kept
alive until they arrive at the destination.
Some requirements should be fulfilled in order to
keep the fish (in this case grouper fish) alive such
as: The density of the fish per m
3
of water in the
fish hold, cleanliness of the water in the fish hold,
salinity of the water, water temperature, mineral
content and concentration in the water, oxygen
and nitrogen contain, and the water circulation
speed.
Since the vessel was designed as a trimaran, all
the propulsion machinery are installed in the side
hulls, and the middle hull is dedicated for the
storing of fish, the water and air circulation system
is arranged in the middle body of the middle hull.
Two pumps are utilized one for circulation fresh
sea water and the other for discharging the
contaminated or dirty water. Air is blown in from
a compressor through distribution pipes to all the
fish holds. For keeping the right temperature in
the fish hold sea water is sucked from the sea chest
and passed through a chiller prior being flown into
the fish hold.
With 16.3 ton fish hold capacity it is estimated
that 0.38 ton of live grouper fish could be
transported.
Key words - trimaran, fishing vessel, fish hold, water
and air circulation
I. INTRODUCTION
Indonesia has abundant of natural resources
including fish and aquaculture, but so far the effort for
exploiting it has not optimal yet and the traditional
fishermen are still using traditional ways of fishing
methods. In order to improve the fishermens standard
of living various efforts have been done among others
is by introducing fish farming for certain types of fish
that have prospective in global market, grouper is one
of such kind of fish. But grouper has to be exported
live, and based on this requirement the research is
focused on the method for transporting the live fish.
The type of boat for carrying the fish is in the form of
trimaran boat, which is considered more stable
compared to the mono hull. In order to keep the fish
alive the water and air circulation in the fish hold
should be well maintained, this include the density of
the fish per m
3
of water in the fish hold, cleanliness of
the water in the fish hold, salinity of the water, water
temperature, mineral content and concentration in the
water, oxygen and nitrogen contain, and the water
circulation speed.
The objectives of the research are: to introduce
the novel design of trimaran vessel, which has its
propulsion engines installed in the side hulls, and the
middle hull is fully used for carrying the live fish, to
design the water and air circulation system in the fish
holds.
II. ARRANGEMENT OF THE BOAT
Grouper has the following requirements in order
to keep them alive:
1. Salinity : 30 - 34 ppt (part per thousand) or
permil,
2. Water temperature : 24 -32 degree Celcius,
3. Water PH :7 9,
4. Speed of water flow : 20 50 cm/second
5. Content of oxygen solution minimum 3 ppm,
6. Nitrogen nitrite : 0 - 0.05 ppm,
7. Ammonia (NH
3
-N) < 0.02 ppm
8. Dark environment because grouper is night
active fish,
9. Density of fish in the water (kg/m
3
),
10. height of the fish hold, and
11. the wide of the hatch.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-24 2/3
To cater the above requirements the boat is
designed as a trimaran boat with the following
specificationsthe side hull are the smaller parts of the
boat where the propulsion engines are located.
1. Middle hull:
The middle hull is the largest part of the boat
where the fish hold are located, the main particulars of
the middle hull are as follows:
Table 1. Middle hull main particulars
There are three fish holds in the middle hull,
which will be filled with water so that the fish can be
kept alive; one pump room located in the middle
where all the machinery for maintaining the fish hold
conditions is located; above the pump room is the
wheel house. Arrangement of the middle hull can be
seen in the figure 1.
Figure
1.
Middle hull
2. Side hull
The size of the side hulls are smaller than the
middle hull, their purpose is to stabilize the boat, in
both of the side hulls the propulsion engines are
installed. The main particulars of the side hulls are as
follows:
Table 2. Side
hulls main
particulars
Figure 2. Side hulls
The arrangement of the fish hold are as follows:
Table 3. Fish hold division:
The capacity of the boat is 380 kg of live grouper
consists of 636 of fish which have economical value.
The sea water needed for the fish is 15,86 m
3
or
16,255 ton.
III. THE WATER AND AIR CIRCULATION
SYSTEM
The water and air circulation system for the fish
holds is generated by the following equipment:
1. Circulation pump 2 HP.
2. General pump 2 HP.
3. Filter
4. Generator
5. Chiller
6. Compressor
Loa 15 m
Lwl 14,55 m
Lpp 14,2 m
D 1,31 m
H 1,67 m
B 2,84 m
Cb 0,47
Cw 0,647
Cm 0,968
Cp 0,486
Loa 10 m
Lwl 9,7 m
Lpp 9,46 m
D 0,87 m.
H 1,12 m.
B 1,9 m
Cb 0,32
Cw 0,55
Cm 0,97
Cp 0,33
Item Hold
1
Hold
2
Hold
3
total
Volume
(m
3
)
2,45 7,5 5,91 15,86
Berat
(ton)
2,51
1
7,687 6,057 16,25
5
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-24 3/3
7. circulation pipes
8. suction pipes for dirty water
IV. CONCLUSIONS
After being calculated for the proposed trimaran boat
the water and air circulation system can be concluded
as follows:
1. The total fish hold capacity is 16,225 ton,
consisted of 0,38 ton grouper live fish, and
15,875 ton of sea water.
2. There are two pumps being used one for
circulating the water and general pump for
pumping the dirty water, each is powered by
2 hp electric motor.
3. To maintain the water temperature a chiller is
being used.
REFERENCES
1. Geankoplis,C.J. Transport Processes and Unit
Operations (New Delhi: Prentice Hall of India, 1997).
2. J.O, Traung, Fishing Boat Of The World 2 ( England:
Fishing New Books Ltd,
1960), hal 87 89
3. Kern, D.Q., Process Heat Transfer (Tokyo: International
Student Edition, Mc Graw Hill Book Co, 1965).
4. Mallawa. A, Sudirman, Teknik Penangkapan Ikan
( Jakarta: PT. Rineka Cipta, 2004).
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-25 1/5
Local Protein Structure Comparison
1
Ford Lumban Gaol,
2
Belawati Widjaja
1,2
Faculty of Computer Science, University of Indonesia,
Kampus UI Depok, Jawa Barat, Indonesia
Tel. 021-7863419, fax. 021-7863415 email : flg50@ui.ac.id
Abstract Genome sequencing projects are working
to determine the complete genome sequence for
several organisms. The sequencing projects have
produced significant impact on bioinformatics
research by stimulating the development of
sequence analysis tools such as methods to identify
genes in a genome sequence, methods to predict
alternative splicing sites for genes, methods that
compute the sequence homology among genes, and
methods that study the evolutionary relation of
genes, to name a few.
Proteins are the products of genes and the building
blocks for biological function. In this paper, we
identify some basic background on local proteins
structure, protein function and their simulation.
Keywords Genome sequence, Pattern-Based
Structure Alignment, Functional Site Identification,
Structure-based Functional Annotation.
I. PROTEIN STRUCTURE
1.1 Amino Acids
Proteins are chains of -amino acid molecules. An
-amino acid (or simply an amino acid) is a molecule
with three chemical groups and a hydrogen atom
covalently bonded to the same carbon atom, the C
atom. These groups are: a carboxyl group (-COOH),
an amino group (-NH2), and a side chain with
variable size (symbolized as R) [BT91]. The first
carbon atom in a side chain (one that is connected to
the C atom) is the C atom and the second one is the
C atom and so forth. Figure 1 illustrates an example
of amino acids.
Different amino acids have different side chains.
There are a total of 20 amino acids found in naturally
occurring proteins. At physiological temperatures in a
solvent environment, proteins adopt stable three-
dimensional (3D) organizations of amino acid
residues that are critical to their function.
1.2 Four Levels of Protein Structure
The levels are as follows:
Figure 1: Up: a schematic
illustration of an amino acid.
Down: the 3D structure
of an amino acid
(Alanine) whose side
chain contains a single
carbon atom. The atom
types are shown;
unlabeled atoms are
hydrogens. The schematic
diagram is adopted from [BT91] and the 3D structure is
drawn with the VMD software.
Primary structure describes the amino acid
sequence of a protein.
Secondary structure describes the pattern of
hydrogen bonding between amino acids
along the primary sequence. There are three
common types of secondary structures: -
helix, -sheet, and turn.
Tertiary (3D) structure describes the
protein in terms of the coordinates of all of
its atoms.
Quaternary structure applies only to
proteins that have at least two amino acid
chains. Each chain in a multi-chain protein is
a subunit of the protein and the spatial
organization of the subunits of a protein is
the quaternary structure of the protein. A
single-subunit protein does not have a
quaternary structure.
Primary Structure
In a protein, two amino acids are connected by a
peptide bond, a covalent bond formed between the
carboxyl group of one amino acid and the amino
group of the other with elimination of a water
molecule. After the condensation, an amino acid
becomes a amino acid residue (or just a residue, for
short). The C
=
(4)
n = number of parameter
The Water Quality Index uses a scale from 0 to 100 to
rate yhe quality of the water, with 100 being the highest
possible score. Once the overall WQI score is known, it
can be compared against the following scale to
determine how healthy or how pollute the water is on
giver time. Water quality cab be scale:
a. 91-100, excellent water quality
b. 71-90, good water quality
c. 51-70, average water quality
d. 26-50, fair water quality
e. 0-25, poor water quality
Raw water quality with rating falling in the excellent
and good range would able to support a high diversity
of aquatic life. In addition, the water would also be
suitable for all forms of recreation, including those
involving direct contact with water. Raw water quality
achieving only an average rating generally have less
diversity of aquatic organisms and frequently have
increased algae growth. Raw water quality falling into
the fair range are only be able to support a limited
number of aquatic life forms, and it is expected that
these water have abundant quality problem. A raw
water quality with poor quality rating would not
normally be considered acceptable for activities
involving direct contact with the water, such as
swimming.
The scale with value 100, will be based on maximum
permissible as mention on Government Decree Number
82-2001 for class 1 of raw quality. The scale with
value 0 will base on class 4 of raw water quality.
Raw water quality parameter will calculate from
dominant parameter occurred during time of
observation. The weight factor will use correlation from
statistic.
III. EXPERIMENTAL RESULTS
By using multivariate analysis among number raw
water parameter observed there are 9 parameter the
most significance parameter among 48 parameter
observed regularly by water supply enterprises. The 9
parameter are , turbidity, Total Dissolved Solid,
ammonia, iron, manganese, BOD, COD, suspended
solid and total coli form.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-33 3/7
From Table 2, as can be seen raw water quality
comparing with water quality standard adopted from
Governor DKI Decree and Government Decree.
Tabel 2. Raw Water Quality Status (Year 2003-
2006)
Raw Water Quality
(Maximum)
No
.
Para
meter
Unit
SK Gub.DKI
No.582/
1995
PP No.
82/2001
B P 1 P 2 PG
1. Turbidity NTU 100 - 15326 8800 9060 13246
2. TDS mg/l 500 1000 312 330 282 241
3. Ammonia mg/l 1,0 0,5 2,28 2,62 2,75 4
4. Iron mg/l 2,0 0,3 13,3 0,95 0,88 29,8
5. Manganese mg/l 0,5 0,1 6,79 5,5 1,48 5,6
6. BOD mg/l 10 2 60 42 41 60
7. COD mg/l 20 10 106 144 101 210
8. S S mg/l 100 50 5290 2834 2540 14560
9. Total coli
x 10
Nol/
100 ml
10 1 201 4800 680 201
Sources: Observation and calculation (2006)
From above table, all of parameter exceed the quality
standard, and the highest number occurred will disturb
water supply treatment operation. The result of
calculation each method can be seen on Table 3, Table
4 and Table 5.
Table 3. Calculation Water Quality Method Storet
(Year 2003-2006)
Year
Raw
Water
Buaran
Raw Water
Pulo
Gadung
Raw Water
Pejompongan
I
Raw Water
Pejompongan
II
2003 -72 -76 -64 -58
2004 -74 -74 -66 -66
2005 -72 -74 -64 -66
Source: Calculation (2006)
From above table can be concluded that during year
2003 until 2006 raw water quality show in heavy
pollution condition ( > -31) as a result strong effort
made in produce water supply with quality standard
given
Table 4. Calculation Water Quality Method Pollution
Index
(Year 2003-2006)
Year
Raw
Water
Buaran
Raw Water
Pulo
Gadung
Raw Water
Pejompongan
I
Raw Water
Pejompongan
II
2003 11,75 11,79 10,91 10,35
2004 10,10 9,90 10,54 9,99
2005 10,70 10,34 11,09 11.03
Source: Calculation
From above table can be concluded that during year
2003 until 2006 raw water quality show in heavy
pollution condition ( > 10) as a result strong effort
made in produce water supply with quality standard
given or treatment process stop.
Table 5. Calculation Water Quality Method WQI
(Year 2003-2006)
Year
Raw
Water
Buaran
Raw Water
Pulo
Gadung
Raw Water
Pejompongan
I
Raw Water
Pejompongan
II
2003 21,21 21,21 22,02 25,08
2004 27,8 15,01 34,46 32,20
2005 15,98 15,01 45,22 46,88
Source: Calculation
From above table, almost all of raw water quality can
be categorize as the worst raw water quality and
indication have less diversity of aquatic organisms and
frequently have increased algae growth.
With that condition, water treatment plan is not allow to
operate. Since, water supply is very essential for human
being, follow up action has done and as a result more
chemical put to treat the water.
Water quality in Water Treatment Pejompongan better
than other water intake, since in the distribution gate in
Cawang , there is some addition treatment, by using
screen, turbulence as a result on water gate and pump
and additional chlorination put into transmission line.
IV. CONCLUSIONS
Population growth and increasing industrial along the
river which are polluting the river and erosion caused
by rain fall made raw water for water supply from
West Tarum Channel become worst. Special treatment
by adding more chemical in processing the water has
already done as a result more sludge produced. In the
near future, modification on water treatment plan is
needed so that water treatment plan can treat more
efficiency .
Many regulation in protecting raw water has made,
however in implementing that regulation need more
serious and coordination in between institution
relevance should be made.
REFERENCES
[1] American Water Works Association, 1999. Water
Quality and Treatment, A Handbook of
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-33 4/7
Community Water Supplies, fifth edition,
McGraw Hill, 1999.
.
[2] Bordalo A.A, Nilsumranchit W and Chalermwat K,
2001. Water Quality and Uses of the Bangpakong
River (Estern Thailand), Water Research Volume
35 No. 15 (3635-3642). 2001.
Appendices
Appendices 1. Calculation on Storet Method (2003)
Para
met
er
U
ni
t
PP
82
/20
01
Raw Water Quality
Buaran
Pulo
Gadung
Pejompong
an 1
Pejompong
an 2
Quali
ty
V
al
u
e
Quali
ty
V
al
u
e
Quali
ty
V
al
u
e
Quali
ty
V
al
u
e
Turb
idity
N
T
U
100
1070
8
-
1
1176
8
-
1 8089
-
1 5808
-
1
76
-
1 150
-
1 135
-
1 169
-
1
4295.
57
-
3
3968.
29
-
3
2983.
43
-
3 2659
-
3
TDS m
g
/l
500
288 0 241 0 330 0 282 0
92 0
145.
8 0 138 0
139.
1 0
170.
93 0
193.
04 0
199.
19 0
185.
05 0
Am
moni
a
m
g
0.5
2.28
-
2 4
-
2 2.62
-
2 2.75
-
2
0.31 0 0.89
-
2 0.25 0 0.24 0
1.03
-
6 2.09
-
6 0.98
-
6 1.00
-
6
Iron m
g
/l
0.3
12.6
0
-
2
19.2
4
-
2 0.95
-
2 0.46
-
2
1.17
-
2 3.87
-
2 0.03 0 0.03 0
5.51
-
6 9.88
-
6 0.22 0 0.14 0
Man
gann
ese
m
g
/l
0.1
6.80
-
2 4.03
-
2 0.38
-
2 0.29
-
2
0.09 0 0.12
-
2 0.03 0 0.01 0
1.53
-
6 1.04
-
6 0.15
-
6 0.10 0
BOD m
g
/l
0.1
60
-
2 60
-
2 42
-
2 35
-
2
8
-
2 7
-
2 12
-
2 14
-
2
18.8
6
-
6 18.5
-
6
23.1
4
-
6
21.6
4
-
6
COD m
g
/l
20
78.2
-
2 78.2
-
2 101
-
2 101
-
2
8.1 0 9 0 20 0 20 0
28.8
4
-
6
29.8
4
-
6
42.7
1
-
6
44.2
1
-
6
SS m
g
100
5180
-
2 5180
-
2 675
-
2 1107
-
2
35 0 19 0 105 0 117 0
1546
-
6
1077
.43
-
6
311.
14
-
6
352.
14
-
6
Total
Coli
m
g
/l
100
00 2010
00
-
3
2010
00
-
3
4200
000
-
3
5600
00
-
3
7000
-
3
7700
0
-
3
2800
00
-
3 1800
-
3
15342
9 -9
17928
6 -9
10200
00 -9
17482
9
-
9
Total
-
7
2
-
7
6
-
6
4
-
5
8
Appendices 2. Calculation on Storet Method (2004)
Paramete
r Unit
PP
82/
Raw Water Quality
2001 Buaran
Pulo
Gadung
Pejomponga
n 1
Pejompongan 2
Quali
ty
V
al
u
e
Quali
ty
V
al
u
e
Quali
ty
v
al
u
e
Quali
ty
Valu
e
Turbidity NTU 100
1532
6
-
1 9242
-
1 6930
-
1 5730 -1
70
-
1 138
-
1 116
-
1 101 -1
5140.1
9
-
3
3830.7
1
-
3
3043.1
4
-
3 2771 -3
TDS mg/l 500
312 0 208 0 292 0 256 0
103 0
121.
4 0 107 0 122 0
157.
82 0
156.
46 0
184.
79 0
179.
2 0
Ammonia mg 0.5
1.21
-
2 1.75
-
2 1.67
-
2 1.21 -2
0.34 0 0.21 0 0.36 0 0.26 0
0.73
-
6 0.98
-
6 0.99
-
6 0.79 -6
Iron mg/l 0.3
12.9
59
-
2
29.7
6
-
2 0.16 0 0.38 -2
0.93
1
-
2 2.27
-
2 0.02 0 0.02 0
7.97
-
6
14.4
3
-
6 0.08 0 0.12 0
Mangannes
e
mg/l 0.1
2.77
-
2 5.6
-
2 0.90
-
2 1.48 -2
0.15
4
-
2 0.21
-
2 0.17
-
2 0.01 0
0.85
-
6 1.60
-
6 0.43
-
6 0.45 -6
BOD mg/l 0.1
18
-
2 14
-
2 34
-
2 24 -2
8
-
2 8
-
2 12
-
2 12 -2
11.9
3
-
6
11.5
7
-
6
20.0
7
-
6
17.5
7 -6
COD mg/l 20
34
-
2 160
-
2 82
-
2 96 -2
9.6 0 16.5 0 16 0 18 0
26.7
-
6
55.7
9
-
6
38.7
1
-
6
36.1
4 -6
SS mg 100
5290
-
2 6210
-
2 2834
-
2 2378 -2
26 0 40 0 109
-
2 113 -2
1873.6
1
-
6
1437
.5
-
6
738.
29
-
6
622.
64 -6
Total Coli mg/l 1000
0
2010
00
-
3
2010
00
-
3
4800
000
-
3
5000
00 -3
1550
00
-
3
2010
00
-
3
1300
0
-
3 100 -3
194428.
57
-9
20100
0 -9
13635
71 -9
10308
0 -9
Total
-
7
4
-
7
4
-
6
6 -66
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-33 5/7
Appendix 3. Calculation on Storet Method (2005)
Para
meter
Uni
t
PP
82
/
Raw Water Quality
200
1 Buaran Pulo Gadung Pejompongan 1 Pejompongan 2
Quality
Va
lue Quality
Va
lue Quality
Va
lu
e Quality
Va
lue
Turbidit
y
NTU 10
0
11068 -1 13146 -1 8800 -1 9060 -1
1333 -1 2754 -1 49.1 -1 1504 -1
6750.4
3 -3 7042 -3
4598.8
7 -3 5684.36 -3
TDS mg/
l
50
0
179 0 192 0 207 0 179 0
91 0 121 0 107 0 115 0
128.93 0 154.31 0 153.71 0 149.21 0
Amonia mg 0.5
1.28 -2 2.52 -2 1.28 -2 1 -2
0.48 0 0.3 0 0.21 0 0.3 0
0.84 -6 1.21 -6 0.52 -6 0.69 -6
Besi mg/
l
0.3
9.313 -2 24.53 -2 0.06 0 0.18 0
3.696 -2 5.25 -2 0.02 0 0.02 0
7.36 -6 13.61 -6 0.04 0 0.08 0
Mangan mg/
l
0.1
1.103 -2 2.31 -2 0.79 -2 0.62 -2
0.091 0 0.23 -2 0.17 -2 0.17 -2
0.56 -6 0.73 -6 0.42 -6 0.34 -6
BOD mg/
l
0.1
13 -2 20 -2 42 -2 41 -2
9 -2 8 -2 12 -2 9 -2
11 -6 12.64 -6 21.5 -6 19.86 -6
COD mg/
l
20
106 -2 210 -2 61 -2 52 -2
5.6 0 14 0 20 0 20 0
48.56 -6 89.64 -6 36.43 -6 35.43 -6
SS mg 10
0
4590 -2 14560 -2 1019 -2 1101 -2
71 0 85 0 131 -2 166 -2
1803.5 -6 2815.5 -6 569.29 -6 556.29 -6
Total
Coli
mg/
l
10
00
0 201000 -3 201000 -3
480000
0 -3 390000 -3
201000 -3 62000 -3 440000 -3 260 -3
201000 -9 177857 -9
147642
9 -9
86972.8
6 -9
Jumlah -72 -74
-
66 -66
Appendix 4. Calculation on Pollution Index (2003)
Par
a
me
ter
Un
it
PP
82
20
01
Kualitas Air Baku (Maksimum, Minimum, Rata-rata)
Buaran Pulo Gadung Pejompongan 1 Pejompongan 2
Li Ci Ci/
Li
Ci/
Li
ba
ru
Ci Ci/
Li
Ci/
Li
ba
ru
Ci Ci/
Li
Ci/
Li
ba
ru
Ci Ci/
Li
Ci/
Li
ba
ru
Tur
bidi
ty
NT
U
10
0
10
70
8
10
7.
1
11
.1
48
11
76
8
11
7.
7
11
.3
5
80
89
80
.8
9
10
.5
3
58
08
58.
08
9.
82
42
96
42
.9
6
9.
16
5
39
68
.2
9
39
.6
8
8.
99
29
83
.4
3
29
.8
3
8.
37
26
59
26.
59
8.
12
TD
S
m
g/l
50
0
28
8
0.
58
0.
58
24
1
0.
48
0.
48
33
0
0.
66
0.
66
28
2
0.5
6
0.
56
17
1
0.
34
0.
14
19
3.
04
0.
39
0.
39
19
9.
19
0.
40
0.
4
18
5.
05
0.3
7
0.
37
Am
mo
nia
m
g
0.
5
2.
28
4.
56
4.
29
4 8 5.
5
2.
62
5.
24
4.
59
2.
75
5.5 4.
7
1.
03
2.
06
2.
56
9
2.
09
4.
17
4.
1
0.
98
1.
97
2.
47
1.
00
2.0
1
2.
51
Iro
n
m
g/l
0.
3
12
.6
04
42
.0
1
9.
11
19
.2
4
64
.1
3
10
.0
3
0.
95
3.
17
3.
5
0.
46
1.5
3
1.
92
5.
51
18
.3
7
7.
32
9.
88
32
.9
3
8.
58
0.
22
0.
73
0.
73
0.
14
0.4
5
0.
45
Ma
ng
an
ese
m
g/l
0.
1
7 67
.9
5
10
.1
6
4.
03
40
.3
1
9.
02
0.
38
3.
77
3.
88
0.
29
2.8
9
3.
3
1.
53
15
.3
6.
92
1.
04
10
.4
5
6.
09
0.
15
1.
53
1.
92
0.
10
1.0
2
1.
04
BO
D
m
g/l
0.
1
60 60
0
14
.8
9
60 60
0
14
.8
9
42 42
0
14
.1
1
35 35
0
13
.7
2
19 19
0
12
.3
9
18
.5
18
5
12
.3
3
23
.1
4
23
1.
43
12
.8
2
21
.6
4
21
6.4
3
12
.6
7
CO
D
m
g/l
20 78
.2
3.
91
3.
96
78
.2
3.
91
3.
96
10
1
5.
05
4.
51
10
1
5.0
5
4.
51
28
.8
1.
44
1.
79
29
.8
4
1.
49
1.
86
42
.7
1
2.
14
2.
65
44
.2
1
2.2
1
2.
72
SS m
g
10
0
51
80
51
.8
9.
57
51
80
51
.8
9.
57
67
5
6.
75
5.
14
11
07
11.
07
6.
22
15
46
15
.4
6
6.
94
10
77
.4
3
10
.7
7
6.
16
31
1.
14
3.
11
3.
46
35
2.
14
3.5
2
3.
73
Tot
al
Col
i
m
g/l
10
00
0
20
10
00
20
.1
7.
51
20
10
00
20
.1
7.
51
42
00
00
0
42
0
14
.1
1
56
00
00
0.0
03
0.
00
3
15
34
29
15
.3
4
6.
92
17
92
86
17
.9
3
7.
26
10
20
00
0
10
2
11
.0
4
17
48
29
17.
48
7.
21
Ma
ksi
mu
m
14
.8
9
14
.8
9
14
.1
1
13
.7
2
Su
m
7.
38
7.
52
6.
26
5.
12
PIy 11
.7
5
11
.7
9
10
.9
1
10
.3
5
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-33 6/7
Appendix 5. Calculation on Pollution Index (2004)
Para
me
Ter
Unit PP
82
20
01
Kualitas Air Baku (Maksimum, Minimum, Rata-rata)
Buaran Pulo Gadung Pejompongan 1 Pejompongan 2
Li Ci Ci/L
i
Ci/
Li
bar
u
Ci Ci/
Li
Ci/
Li
bar
u
Ci Ci/L
i
Ci/L
i
bar
u
Ci Ci/L
i
Ci/L
i
bar
u
Turbi
dity
NTU 10
0
153
26
153
.3
11.
92
924
2
92.
42
10.
82
693
0
69.
3
10.
2
573
0
57.
3
9.7
9
514
0
51.
4
9.5
5
383
0.7
1
38.
31
8.9
1
304
3.1
4
30.
43
8.4
1
277
1
27.
71
8.2
1
TDS mg/l 50
0
312 0.6
2
0.6
2
208 0.4
2
0,4
2
292 0.5
8
0.5
8
256 0.5
1
0.5
1
158 0.3
2
0.3
2
156
.46
0.3
1
0.3
1
184
.79
0.3
7
0.3
7
179
.2
0.3
6
0.3
6
Amm
onia
mg 0.5 1.2
1
0,0
2
0.0
2
1.7
5
3.5 3.7
2
1.6
7
3.3
4
3.6
1
1.2
1
2.4
2
2.9
1
0.7
3
1,4
6
1.8
2
0.9
8
1.9
5
2.4
5
0.9
9
1.9
8
2.4
8
0.7
9
1.5
9
2
Iron mg/l 0.3 13 43.
20
9.1
7
29.
76
99.
19
10.
98
0.1
6
0.5
3
0.5
3
0.3
8
1.2
7
1.5
2
7.9
7
26.
57
8.1
2
14.
43
48.
09
9.4
1
0.0
8
0.2
5
0.2
5
0.1
2
0.4
1
0.4
1
Man
gane
se
mg/l 0.1 2.7
7
27.
7
8.2
1
5.6 56 9.7
4
0.9
0
8.9
6
5.7
6
1.4
8
14.
8
6.8
5
0.8
5
8.5 5.6
4
1.6
0
16.
00
7.0
2
0.4
3
4.2
7
4.1
5
0.4
5
4.4
6
4.2
4
BOD mg/l 0.1 18 180 12.
27
14 140 11.
7
34 340 13.
65
24 240 12.
9
12 120 11.
39
11.
57
115
.7
11.
31
20.
07
200
.7
12.
5
17.
57
175
.71
12.
22
COD mg/l 20 34 1.7 2.1
5
160 8 5.5
1
82 4.1 4.0
6
96 4.8 4.4
16.
7
0.8
4
0.8
4
55.
79
2.7
9
3.2
2
38.
71
1.9
4
2.4
3
36.
14
1.8
1
2.2
9
SS mg 10
0
529
0
52.
9
9.6
1
621
0
62.
1
9.9
6
283
4
28.
34
8.2
6
237
8
23.
78
7.8
8
187
4
18.
74
7.3
6
143
7.5
14.
38
6.7
8
738
.29
7.3
8
5.3
4
622
.64
6.2
3
4.9
7
Total
Coli
mg/l 10
00
0
2E+
06
201 12.
5
201
000
20.
1
7.5
1
480
000
0
480 14,
4
500
000
50 9.4
9
194
429
19.
44
7.4
4
201
000
20.
1
7.5
1
136
357
1
136
.4
11.
67
103
080
10.
31
6.0
6
Maks
imu
m
12.
5
11.
7
13.
65
12.
9
Sum 6.9
2
7.7
0
5.9
9
5.7
8
PIy 10.
1
9.9 10.
54
9.9
9
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-33 7/7
Appendix 6. Calculation on Pollution Index (2005)
Parame
ter
Satuan PP 82
2001
Kualitas Air Baku (Maksimum, Rata-rata)
Buaran Pulo Gadung Pejompongan 1 Pejompongan 2
Li Ci Ci/Li Ci/Li
baru
Ci Ci/
Li
Ci/
Li
ba
ru
Ci Ci/
Li
Ci/
Li
ba
ru
Ci Ci/L
i
Ci/
Li
ba
ru
Turbidity NTU 100 11068 110.7 11.22 131
46
131
.5
11.
59
880
0
88 10.
72
906
0
90.
6
10.
78
6750 67.5 10.14 704
2
70.
42
10.
23
459
8.8
7
45.
99
9.3
1
568
4.4
56.
84
9.7
7
TDS mg/l 500 179 0.36 0.36 192 0.3
8
0.3
8
207 0.4
1
0.4
1
179 0.3
6
0.3
6
128.9 0.26 0.26 154
.31
0.3
1
0.3
1
153
.71
0.3
1
0.3
1
149
.21
0.3
0
0.3
Ammonia mg 0.5 1.28 2.56 3.04 2.5
2
5.0
4
4.5
1
1.2
8
2.5
6
3.0
4
1 2 2.5
0.84 1.68 1.13 1.2
1
2.4
2
2.9
2
0.5
2
1.0
4
1.0
8
0.6
921
1.3
8
1.6
9
Iron mg/l 0.3 9 306.7 13.43 24.
53
81.
76
10.
56
0.0
6
0.2 0.2 0.1
8
306
.67
13.
42
7.36 24.53 7.94 13.
61
45.
38
9.2
8
0.0
4
0.1
2
0.1
2
0.0
785
0.2
6
0.2
6
Mangan mg/l 0.1 1 11.03 6.21 2.3
1
23.
13
7.8
2
0.7
9
7.8
5
5.4
7
0.6
2
6.2 4.9
6
0.56 5.6 4.74 0.7
3
7.3
0
5.3
1
0.4
2
4.1
6
4.0
9
0.3
419
3.4
2
3.6
7
BOD mg/l 0.1 13 130 11.57 20 200 12.
5
42 420 14.
11
41 410 14.
06
11 110 11.2 12.
64
126
.43
11.
5
21.
5
215 12.
66
19.
857
215 12.
66
COD mg/l 20 106 5.3 4.62 210 10.
5
6.1 61 3.0
5
3.4
2
52 3.0
5
3.4
2
48.6 2.43 2.92 89.
64
4.4
8
4.2
5
36.
43
1.8
2
2.3 35.
429
1.8
2
2.3
SS mg 100 4590 45.9 9.3 145
60
145
.6
11.
81
101
9
10.
19
6 110
1
10.
19
6
1803 18.03 6.28 281
5.5
28.
16
8.2
5
569
.29
5.6
9
4.7
7
556
.29
5.6
9
4.7
7
Total Coli mg/l 10000 201000 20.1
7.51
201
000
20.
1
7.5
2
480
000
0
480 14.
4
390
000
480 14.
4
201000 20.1 7.51 177
857
17.
79
7.2
5
147
642
9
147
.6
11.
84
869
73
147
.64
11.
85
Maksimum 13.43 12.
5
14.
4
14.
4
Sum 6.99 7.6
1
6.2
4
6.9
2
PIy 10.7 10.
34
11.
09
11,
03
Non Linear Dynamic Identification of Waste Heat
Boiler Unit Based on
Adaptive Neuro Fuzzy Inference System (ANFIS)
Yuliati
Electrical Engineering Department, Widya Mandala Surabaya Catholic University.
Jl. Kalijudan 37 Surabaya,
Phone: +62-31-3891264, Email : yuliatheresia@yahoo.com
Abstract- Waste Heat Boiler (WHB) unit is a
crucial plant for a fertilizer plant especially for
steam product .The water level in the its steam
drum shall be maintained at certain limit in order
to give a good steam quality. It is generally
accepted that a boiler unit is a highly nonlinear
and strongly coupled complex system. Due to the
dynamic characteristic, it is difficult to have its
model in order to give suitable control system of
the steam drum water level theoretically.
The success of model-based non linear control
technique is usually conditioned by the
availability of accurate models which reflects the
non linear process complexities. Prior to
implementing any control scheme, its dynamic
properties need to be identified properly. A real
time data driven neuro fuzzy model will be
presented for the water level of boiler unit. Neuro
fuzzy design combine architectural of neural
network and philosophical of fuzzy system aspects
of an expert resulting in an artificial brain, which
can be used as an identifier.
An experiment for identifying the water level
of boiler unit was conducted on real-time plant
operation of a fertilizer plant in Gresik, East
Java, Indonesia. Validation model was also
performed to verify the results of identification.
The performance of the procedure shows how the
neuro fuzzy algorithm produces accurate models
for non linear control design.
Keywords : system identification, waste heat boiler,
real plant, fuzzy logic, neural network.
I. Introduction
There are two types of boilers in the steam
generation i.e. fired boilers and unfired boilers (heat
recovery steam generator) is in fact a heat exchanger
extracting heat from through put gases and raising
water to steam condition. Waste heat boiler unit
representing type of unfired boiler system which
using heat transfer energy form gas turbine exhaust
gas flow. Its about 60% of 70% of the heat that
obtained by combustion in gas turbine, re-exploited
by WHB to produce the steam.
Its generally accepted that a boiler unit is a
highly non linear and strongly coupled complex
system. Its dynamic characteristic makes it very
difficult to model and control the water level of
steam drum theoretically. The water level of High
pressure steam drum in WHB unit should be
maintained within safe limit. A too high of water
level produces wet steam that could damage the
turbine equipment and the turbine can trip off. On the
other hand, too low of water level risks uncovering
the water tubes and exposing them to heat stress and
damage. In both cases, the plant shuts down
unintentionally.
The success of model-based non linear control
technique is usually conditioned by the availability of
accurate models which reflects the non linear process
complexities. Prior to implementing any control
scheme, its dynamic properties need to be identified
properly.
Neural network have been applied successfully
in the identification and control of non linear
dynamic system. The universal approximation
capabilities of multilayer perceptron (MLP) make it
popular choice for modelling non linear systems.
However, neural network architecture is rather
complicated and difficult to be understood by human.
On the other hand, the fuzzy inference system can
initialise and learn linguistic and semi linguistic
(Sugeno) rules therefore it can be considered as
direct transfer knowledge which is the main
advantage of fuzzy inference systems but the
learning capabilities of fuzzy system is not as good
as neural network. It is then to develop to integrate
the strength of both methodologies in order to
achieve the learning capabilities and transfer
knowledge via fuzzy if-then rules, producing an
alternative approach so called neuro fuzzy.
In this paper, a development of non linear
identification for high pressure drum (HP drum) of
boiler unit will be design using neuro fuzzy
approach which the fuzzy inference system can be
tuned with a neural network algorithm during
learning based on collection of input output data
obtained from real-time experiment. By the using
real plant, it is expected that the model obtained
could capture the basic feature of water level of
steam drum system and this technique could also
improve the modelling quality of the system.
II. Waste Heat Boiler Plant
Waste Heat Boiler (WHB) unit plays a crucial
role in a fertilizer plant as a part which produces
steam. The temperature of the exhaust gas from gas
turbine is about 600
0
C used to produce superheated
steam which has 65 kg/cm2 pressure and temperature
460
0
C. Boiler Feed Water (BFW) is feed into High
Pressure Steam Drum using 2 pumps which standby
interchange each other. Control valve is used to
control flow of BFW; the level of HP steam drum is
controlled by 2 types of drum level control i.e. single
element and three element control. If the steam
product is smaller then 30 ton/hr, single element
control used to control level drum. In the other hand,
if the steam product is greater then 30 ton/hr, three
elements drum level control used to control level
drum. It combines between level drum, steam flow
and feed water flow.
WHB unit in PT. Petrokimia Gresik is an unfired
boiler completed by some sub system (Figure 1) i.e.
Super heater, high pressure (HP) evaporator, low
pressure (LP) evaporator, economizer, low header,
HP/LP down comers, HP/LP drum, duct burner, and
firing room.
Figure 1. WHB sub system
III. The Experimental Set-up
The experiment set-up is referred to Piping &
Instrumentation Diagram (P&ID) in PT. Petrokimia
Gresik. There are several conditions that must be
considered to design this experiment as follow:
a. Pressure and temperature of BFW,
Pressure and temperature of HP drum,
steam product temperature and blow down
were kept constant as much as possible.
b. Set point flow steam product is based on
the need of steam in urea plant.
c. Data have been collected in real time from
the existing Distributed Control System
(DCS) with sampling time 1 minute.
d. Each 3000 data for every corresponding
inputs and outputs have been collected. The
first 500 data will be used for
identification, and the 1500 remaining data
for model validation.
e. Due to high risk of its plant operation, the
experiment has been conducted by closed
loop manner without change the real plant
operation.
The record of input output data during the conducted
experiment can be shown in Figure 2 and will be
used for modelling and validation purpose.
(a) Flow BFW
(b) WHB Steam Flow
(c) Level WHB HP Drum
Figure 2. Representation of Input-Output Data
During identification process, it can be defined as
shown in Table 1 for input-output variable for
modelling purpose.
Table 1. Input-Output Variable for Modelling
Variable Tag
number
Description
) 1 (
1
k x
FC.22220 Flow BFW (Ton/hr)
) 1 (
2
k x
FC.22221 Steam Product Flow
(Ton/hr)
) 1 ( k y
LC.22220 Level water in HP steam
drum at sampling instant
k-1 as an input model (%)
) (k y
LC.22220 Level water in HP steam
drum at sampling instant k
as an output model (%)
III. Neuro Fuzzy Dynamic Identification
3.1 ANFIS Structure
ANFIS implements a Takagi- Sugeno Fuzzy
Models have to be a powerful tool for the non linear
dynamic identification system. ANFIS scheme as an
estimator model can be shown in Figure 3.
Figure 3. ANFIS as an Estimator Model
The first order of ANFIS structure with three
inputs and one output can be described by the
following rule base:
Rule i
th
: if ) 1 (
1
k x is A
i
and ) 1 (
2
k x is B
i
and
) 1 ( k y is
i
C then
i i i i i
v k y r k x q k x p f + + + = ) 1 ( ) 1 ( ) 1 (
2 1
,
m i ,..., 2 , 1 = (1)
Where m denotes the number of rules. In the other
hand, the non linear AutoRegressive with eXogenous
input (NARX) model is frequently used for non
linear identification techniques. Therefore, in order
to represent the neuro fuzzy for Multi Input Single
Output (MISO) approach can be defined as
)) 1 ( ),..., ( ), 1 ( ),..., ( ( ) 1 ( + + = +
b a m
n k u k u n k y k y f k y
(2)
The input vector x , of the neuro fuzzy model as:
)] 1 ( ),..., ( ), 1 ( ),..., ( [ + + =
b a
n k u k u n k y k y x
(3)
Where ) 1 ( ),..., ( +
a
n k y k y and
) 1 ( ),..., ( +
b
n k u k u are tapped delay lag (TDL)
process output and input respectively. ANFIS has
five layered architecture which has three inputs and
one output can be shown in Figure 4.
Figure 4. Structure of ANFIS
The architecture employs an adaptive network
consist of adaptive nodes in the 1
st
and 4
th
layer. In
the 1
st
is for fuzzification of the input variable and T-
norm operator are deployed in the 2
nd
layer to
compute the rule antecedent part (premise
parameter). The 3
rd
layer normalizes the rule
strengths followed by 4
th
layer where the consequent
parameters are determined. The output network
computes the overall input as the summation of all
incoming signal. The description and computation in
every layer can be described as shown in Table 2.
Table 2. The Description and Computation of Each
Layer in ANFIS Method
Layer Description
1
Every nodes i is an adaptive node with a
node output :
) (
, 1
x O
Ai i
u = , and ) (
2 , 1
y O
Bi i
= u
The output of this layer is the
membership values of the premise part.
The generalized bell is specified by
three parameters {a,b,c}:
i
b
i
i
Ai
a
c x
x
|
|
\
|
+
=
2
1
1
) ( u
has been chosen as membership function
2
Fixed node labelled which multiplies
the incoming signal and output product
represents the firing strength of a rule:
) ( ) (
, 2
y x w O
Bi Ai i i
u u = = , i=1,2
3
Fixed node labelled N which
normalizes the rule strengths
2 1
, 3
w w
w
w O
i
i i
+
= = , i=1,2
4 Every node i is an adaptive node with a
node function :
( )
i i i i i i i
r y q x p w f w O + + = =
, 4
, i= 1,2
Referred to as consequent parameters.
5 This single node is a fixed node which
computes the overall output as the
summation of all incoming signal:
= = =
i
i
i
i
i i
i i i
w
f w
f w f O
5 ,
3.2 Hybrid Learning Algorithm
The hybrid learning algorithm has two passes for
ANFIS i.e. forward pass and backward pass. Suppose
that S1 is a set of non linear parameters and S2 is a
set of linear parameters in the forward pass, node
outputs go forward until 4 layer and the consequent
parameters are identified by the least square
estimation. In the backward pass, the error signals
propagate backward and then premise parameters are
update by gradient descent.
a. Forward pass
If the values of the premise parameters are fixed,
then the overall output can be expressed as a linear
combination of the consequent parameters, the output
f can rewritten as
2
2 1
2
1
2 1
1
f
w w
w
f
w w
w
f
+
+
+
=
( ) ( )
2 2 2 2 1 1 1 1
r y q x p w r y q x p w + + + + + =
( ) ( ) ( ) ( )
2 2 1 1 1 1 1 1
p x w r w q y w p x w + + + =
( ) ( )r w q y w
2 2 2
+ + (4)
From the above equation, it can be seen that the
consequent parameters are linear respect to the
system output. If N learning data is applied to the
equation (4), it can be represented by
y A =
Where unknown parameter in S2 and y is is N
vector data. The problem is how to find the best
solution of this equation that minimize
2
y A ,
is the least square estimator
( ) y A A A
T T
1
*
= (5)
b. Backward pass
In the backward pass the error signal
propagate from the output end toward the input end
using the simple gradient steepest descent for each
training data entry. By minimizing a cost function
defined by
=
=
T
k
k e E
1
2
) (
2
1
(6)
With ) ( ) ( ) ( k y k y k e = (7)
is an error function at time instant k, with y(k) is the
desired response.
For updating the changes of each parameter, the
gradients are needed to minimize E Using the
relation:
) (
) (
) 1 (
) (
k
k E
k k
L
i
L
i
L
i
+ = (8)
Where is the learning rate and
) (
) (
k
k E
L
i
denotes
the order derivative of the error signals (difference
between an actual output and desired output vector)
respect to the node parameter .
IV. The Results of System Identification
Identification process has been conducted using
Multi Input Single Output (MISO) approach as
shown in Figure 4. For learning process and
modelling purpose, assuming that the plant under
consideration can be represented as a first order
process, then the regressor /identifier have been
chosen as
x= )] 1 ( ), 1 ( ), 1 ( [
2 1
k y k x k x (9)
And after 50 epochs for learning process, the ANFIS
structure model can be obtained as the following: the
number of nodes: 34, the number of linear
parameters: 32, the number of nonlinear parameters:
18, the number of training data pairs: 500, and the
number of fuzzy rules: 8.
The result of system identification and its
corresponding errors of learning process can be
shown in Figure 5 and 6, respectively.
Figure 5. The result of System Identification
Figure 6. Error
It can be seen from these figure that the output
of the identified model can follow the measurement
signals satisfactorily and the error (the difference
between these two signals) are relatively small
enough.
V. Model Validation
Beside a direct observation by comparing the
plot of output measurements and the output model
based on the estimated parameter, quantitative
criteria of the Root Mean Square Error (RMSE) can
be used for validation model for a measure the
goodness of fit the model. It can be defined as
RMSE = | |
=
N
t
k y k y
n
1
2
) ( ) (
1
(10)
If the RMSE 5 . 0 then the developed model can be
accepted as a good approach to represent the model
system. The 1500 remaining data was used for
validation purpose, and the result of model validation
and error of validation model can be shown in Figure
7 and 8, respectively.
Figure 7. The Result of Model Validation
Figure 8. Error of the Model Validation
The RMSE both of the identification process and the
model validation are relatively small enough and it
can be seen in Table 3.
Table 3. Root Mean Square Error
Data from :
Identification Validation
RMSE 0.0324 0.0640
VI. Conclusion
An alternative approach to identify level water
of high pressure steam drum in WHB unit using
neuro fuzzy technique has been proposed. By
comparing between the response models with the
actual output measurements revealed that the
satisfactory model matching were obtained. It means
that the non-linear characteristics of system can be
capture from the real basic feature of the WHB
system. The performance of the procedure shows
how the neuro fuzzy algorithm produces accurate
models for non linear control design. Further
application for level control system design based on
identified model is currently under study.
VII. References
[1]. Johanson, R, System Modelling and Identification,
Prentice-Hall International, Inc, NJ, USA, 1993.
[2]. Jyh-Shing Roger Jang, Chuen-Tsai Sun, Neuro Fuzzy
Modelling and Control, The proceeding of the 1995
IEEE, International Conference on Control
Application, 1995.
[3]. J.S.R. Jang, C.T. Sun, E. Mizutani, Neuro Fuzzy and
Soft Computing: A Computational Approach to
Learning and Machine Intelligent, Prentice-Hall,
International Edition, 1997.
[4]. Narendra,K.S, Parthasarathy,K, Identification and
Control of Dynamical Sistems Using Neural Network,
IEEE Transaction on Neural Network, Vol.1, No.1,
March 1990.
[5]. A.W. Ordys, A.W. Pike, M.A. Johnson, R.M. Katebi,
and M.J. Grimble, Modelling and Simulation of
Power Generation Plants, Springer-Verlag, 1994.
[6]. Standard Fasel Lantjes, Waste Heat Boiler, PT.
Petrokimia Gresik, 1992.
[7]. Xiaoguang Chang, A Recurrent Fuzzy- Neuro Dynamic
Model for a Stoked-Fired Boiler, Laboratory of
Intelligent Technology and Systems Tsinghua
University, China, May 1999.
[8]., .Procidia
TM
Control Solutions boiler
Control Overview, Application Data by SIEMENS,
March 2006.
[9]. Yul Y.Nazaruddin, Joko W, S. Hadisupadmo, Inverse
Learning Control using Neuro Fuzzy Approach for a
Process Mini Plant, International Conference on
Physics and Control, Rusia, August 2003.
ISSN: 1411-1284
Proceeding 10
th
Intl QIR 4-6 Dec 2007 EPE-35 1/1
LOAD PATTERN AND REALIBILITY ASSESSMENT OF
POWER GENERATION SYSTEM
Suparman, Zuhal, and Rinaldy D.