Faculty Engineering & Built Environment
Administration
Dean
Prof. Ir. Dr. Mohd Marzuki Mustafa (Until 15th May 2014) Prof. Dato Ir. Dr. Abdul Wahab Mohammad
Deputy Dean (Postgraduate & Development) Prof. Dr. Muhammad Fauzi Mohd Zain
Deputy Dean (Undergraduate & Alumni) Prof. Dr. Ahmad Kamal Ariffin Mohd Ihsan Deputy Dean (Research & Internationalisation) Prof. Ir. Dr. Salina Abdul Samad
Head, Quality Management
Prof. Ir. Dr Mohd Syuhaimi Ab. Rahman Head, Industry & Community Partnerships Prof. Dr. Mohd Zaidi Omar
Head, Department of Civil & Structural Engineering Prof. Dr. Mohd Raihan Taha
Head, Department of Electrical, Electronic & Systems Engineering Prof. Dr. Mahamod Ismail
Head, Department of Chemical & Process Engineering Prof. Dr. Abu Bakar Mohamad
Head, Department of Mechanical & Materials Engineering Prof. Dr. Shahrum Abdullah
Head, Department of Architecture Dr. Nik Lukman Nik Ibrahim
Coordinator for Unit of Fundamental Engineering Studies Assoc. Prof. Dr. Zulkifli Mohd Nopiah
Head, Centre for Automotive Research Assoc. Prof. Dr. Rozli Zulkifli
Head, Centre for Engineering Education Research Prof. Dr. Azami Zaharim
Head, Sustainable Urban Transport Research Centre Prof. Engr. Dr. Amiruddin Ismail
Head, Research Center for Sustainable Process Technology Prof. Dato’ Ir. Dr. Abdul Wahab Mohammad
Senior Assistant Registrar Mr. Zaidi Mohd Noor
The Research Bulletin is published once a year by the Advanced Engineering Center, Faculty of Engineering and Built Environment, UKM. We are pleased to invite Faculty members to contribute their articles to our bulletin. Articles can be submitted:
Advanced Engineering Centre
Faculty of Engineering and Built Environment Universiti Kebangsaan Malaysia
Tel : +603-89118007/8312 Fax : +603-89252546
Email : pghpkt@eng.ukm.my / roslena@eng.ukm.my An electronic version of the Research Bulletin and guidelines to authors is archived at www.ukm.my/pkt
Board of Editorial
Chief Editor
Prof. Ir. Dr. Salina Abdul Samad Editors
Prof. Dr. MA Hannan
Prof. Dr. Siti Kartom Kamarudin
Assoc. Prof. Dr. Mariyam Jameelah Ghazali
Assoc. Prof. Dr. Ahmed Hussein Kamel Ahmed Nasser Elshafie Assoc. Prof. Dr. Zulkifli Mohd Nopiah
Dr. Sudharshan Naidu a/l Raman Mr. Abdul Halim Ismail
Ms. Roslena Binti Md Zaini
Greetings to all readers!
My heartiest congratulations go towards the successful publication of the Seventh Edition of INNOVATE, 2014.
It has been eight years since INNOVATE was first published. It is a publication that is impactful and is seen positively by society towards the Faculty of Engineering & Built Environment (FKAB), and its members’ quest to go global.
Current research, PhD theses abstracts, research projects, awards, achievements, and, the latest, intellectual property recognition are elements found in the bulletin. Ergo, all ideas, feedback and critiques are welcomed to strengthen the faculty’s research potency.
I hope that this bulletin can become FKAB’s conduit in promoting the strength and expertise of the faculty. In addition, I also hope that it becomes the leading platform for FKAB researchers to share their research findings and successes with the university and the world at large.
I truly call upon faculty staff to give their utmost support in this positive endeavor by contributing articles, sharing research triumphs, inventions and innovations so that they become public. By doing this, it will help UKM become a university of great note.
Great thanks go to the editorial board and people who tirelessly contributed to this publication. You are the unsung heroes that ensure this bulletin sees the light of day allowing the nation and the world the pleasure of shared knowledge.
Thank you.
Foreword by Dean
Welcome readers to the Seventh Edition of INNOVATE, research bulletin of the Faculty of Engineering and Built Environment (FKAB) UKM. It is a medium to share research endeavours of members of our faculty and to highlight our achievements for 2013. It is distributed throughout UKM and shared with strategic partners nationwide and internationally.
All feedback is welcome to help us continuously improve in content and impact to our readers. The feedback that we received on the Sixth Edition from our industry partners made us decide to expand the segment on Current Research.
It is gratifying to note that INNOVATE is able to generate interest for the research conducted at our faculty and become a contact platform for our researchers and interested parties outside the university. In this edition, we include more Current Research articles including contributions from the winners of FKAB Research Awards and UKM Innovation Awards. Congratulations and thank you contributors.
Lastly, on behalf of the editorial board, I would like to thank everyone who has worked on this edition of INNOVATE.
Let us, members of the FKAB, continue to share information on our research locally and internationally though INNOVATE and see where it leads us next.
Editorial Note
Prof. Ir. Dr. Salina Abdul Samad
Deputy Dean (Research and Internationalisation) Faculty of Engineering and Built Environment Prof. Dato’ Ir. Dr. Abdul Wahab Mohammad Dean Faculty of Engineering and Built Environment
7
9 10
23
11
12 13 14 15
16 17 18
19 20
25 26 27
28
29 30
31
32
Current Research
•
Wear Failure Analysis of Gas Turbine Combustor Components•
Multiple Security Layers for Protecting Data Using Steganography and Cryptography in Compressed Domain Video•
Application of ANFIS and FIS Systems in Modeling Fatigue Life under Constant and Variable Amplitude Loading•
Potential of Banana Leaf as an Effective Lubricant for High Temperature Application•
Development of Design Method for Thin-Webbed Composite Plate Girders with Partial Interaction•
Deterministic and Probabilistic Finite Element Analysis of Elastic Interacting Cracks•
Design of Dynamic Phase Frequency Detector with Charge Pump and Loop Filter for High Speed PLL•
Nanomaterial for Functional Devices•
Thermal Cooling System for Electric Vehicle Li-Ion Battery•
Intelligent Cooperation Management among Base Stations with Renewable Power System for a Green Fifth Generation (5G) Wireless Network•
Wideband Planar Antenna Design Using Heuristic Optimization Algorithm•
Prototype Implementation of Intelligent Smart Bin for Real Time Solid Waste Status Monitoring•
FPGA Based Human Activity Detection SystemPh.D Thesis 2013 Abstracts
•
Advanced Constitutive Modeling of Granular Aterials Incorporating Fabric Anisotropy Effects•
Adopting Artificial Intelligences in Optimizing Reservoir Operation Policy•
Treatment of Produced Water Using Nanofiltration and Reverse Osmosis Membranes as a Possible Source of Potable Water•
Power Quality Monitor Placement and Voltage Sag Source Location Methods in Power Systems Using Multivariable Regression and Genetic Algorithm•
An Efficient Slow-Start and Congestion Avoidance Mechanisms for Transmission Control Protocol•
A Framework to Assist In Selection and Prioritization of Operational Key Performance Indicators in Engineering Maintenance Organization•
A Real Time Automated Vehicle Incident Detection Using High Accuracy Global Positioning System•
Development of ZnxCd1-xS Window Material for High Efficiency Ultra-Thin Cadmium Telluride (CdTe) Solar CellCONTENTS
INNOVATE | Research Bulletin Engineering & Built Environment
21
•
Improving Quality Management Systems through the Implementation of an Appropriate Quality Culture in Libyan Manufacturing Companies•
Development of a Structural Equation Model to Examine the Impact of Green Supply Chain Management in the Manufacturing Industry•
Progressive Failure Characteristics of Woven Natural Silk Epoxy Composite Tube in Quasi-Static Compression•
Evaluation of a Solar Assisted Combined Ejector-Absorption Refrigeration System with Flash Tank•
La0.6Sr0.4Co0.2Fe0.8O3-d−SDC Carbonate Nano Composite as Cathode for Low Temperature Solid Oxide Fuel Cell•
Fabrication and Charactrisation of Single Button Cells with Sm0.2Ce0.8O1.90 (SDC)/ Y0.25Bi0.75O1.5 (YSB) Bilayer Electrolytes for Intermediate and Low TemperatureSolid Oxide Fuel Cells•
A Framework for Implementing Pre-Development Process in Malaysian Food and Beverages Manufacturing SMEs•
The Implementation of Benchmarking in Malaysian Palm Oil Industry•
The Development of House Defects Index for Terrace House•
Application of Ventilation Optimisation Composite Index (IKOP) in Primary School Offices•
Development of Industrial, Adaptable and Demountable Panel System for Residential LayoutAwards 2013
Achievements 2013
Research Projects 2013
Intellectual Properties 2013
CONTENTS
INNOVATE | Research Bulletin Engineering & Built Environment
45-49 51-55 57-82 83-88
33
34
35
36 38
39
40 41 42 43 44
Current Research
Engineering & Built Environment
99
Gas turbine combustor consists of components designed to burn air/fuel/ mixture and provides a path for directing the hot combustion product to the gas turbine section where mechanical power is produced. In a cannular type combustor (Figure 1), the components are assembled by fitting them together to form a continuous channel. This arrangement is replicated circumferentially around the combustor casing. The method employed for the assembly of the components introduces numerous mating surfaces (Figure 2). Wear is one of the main problems in these components. The mating surfaces of gas turbine combustor components suffered severe wear damages after 8,000 operating hours. A wear failure analysis by using microstructural examination via high resolution secondary electron microscope (SEM) equipped with the energy dispersed x-ray (EDX) is utilized to determine the wear modes and mechanisms.
Wear Failure Analysis of Gas Turbine Combustor Components
The work carried out in this study is a part of a project for assessing degradation of Gas Turbine Combustor Components which is supported by Tenaga Nasional (TNB) research fund. One of the main objectives of this project is to build a systematic failure analysis for combustor components. Thus, a comprehensive study is needed to enhance knowledge and capability for assessing wear failure on gas turbine combustor components.
Associate Professor Dr. Mariyam Jameelah bt Ghazali Department of Mechanical and Material Engineering Faculty of Engineering and Built Environment Universiti Kebangsaan Malaysia
43600 UKM Bangi, Selangor, MALAYSIA
Tel: 603-89216418 Email: mariyam@eng.ukm.my
Figure 1. Detail array of combustor in combustion system Figure 2. Assembly of combustor components
For further correspondence kindly contact :
INNOVATE | Current Research
10
Professor Ir. Dr. Salina Abdul Samad
Department of Electrical, Electronics and Systems Engineering Faculty of Engineering and Built Environment
Universiti Kebangsaan Malaysia
43600 UKM Bangi, Selangor, MALAYSIA Tel: 603-89216331 Email: salina@eng.ukm.my
The rapid intensification of the internet, storage, information systems and digital communications in the past decade has enabled users to send digital data over networks. However, transmission of information in an open network is not secure. Accordingly, developments in both shielding information and communication security have heightened interest in providing more secure techniques for data handling. Steganography, which is defined as imperceptible embedding confidential information within a digital multimedia data host, can be a potential solution to exchange information over the internet without fear of the data being detected. Three important aspects should be considered when using steganography: imperceptibility, capacity, and security. Furthermore, videos distributed via the internet must be processed by a low bit rate compression due to bandwidth limitations. As a result, there is a continual need for innovative solutions that can provide better protection in terms of privacy, integrity, or authentication. One such solution is to combine both steganography and cryptography to increase the level of security and embed private information using techniques such as Embedding Based Byte Differencing (EBBD) during the compression process of host signal. This approach is based on multiple security layers for protecting the data embedded in MPEG-2 compressed video. Data is encrypted and concealed using EBBD in the shuffled content of the host signal using Data Encryption Standard (DES).
The overall scheme is shown in Figure 1. This method has been shown to overcome the drawbacks of current techniques which either is low in payload size or have a noticeable distortion and weakness in security level.
Multiple Security Layers for Protecting Data Using Steganography and Cryptography
Figure 1. Overall scheme for compressed domain video processing For further correspondence kindly contact :
INNOVATE | Current Research
11
Professor Dr. Ahmad Kamal Ariffin Mohd Ihsan.
Department of Mechanical and Materials Engineering Faculty of Engineering and Built Environment
Universiti Kebangsaan Malaysia
43600 UKM Bangi, Selangor, MALAYSIA Tel: 603-89216500 Email: kamal@eng.ukm.my www.eng.ukm.my/cem
Application of ANFIS and FIS Systems in Modeling Fatigue Life under Constant and Variable Amplitude Loading
Fatigue crack growth (FCG) performance of structures is greatly depended on the bulk material properties. It is difficult to identify a fatigue crack in the structures during service. The correct and proper evaluation of fatigue life and fatigue crack propagation is important to ensure the economic consideration, environmental protection and public safety. Hence, this requires a new methodology to predict the crack growth and fatigue life correctly better than current fatigue models widely use. hmThe main objective of this research is to evolve neuro-fuzzy methods to predict the fatigue life and crack growth rate based on a fatigue crack test. Three point bend specimens were used for the experiments, simulation and prediction using ASTM A533 steel alloy according to the ASTM E647-00 standard. For the simulation part, the fatigue models were described, formulated and applied to simulate the fatigue life and FCG rates under constant and variable amplitude loading. The modelling had been carried out based on the modified Forman and NASGRO models using Glyphwork software, which was used to predict FCG after block system construction. Two neuro-fuzzy methods, i.e. ANFIS and CANFIS have been applied to predict the fatigue crack growth rate and fatigue life under the same loading conditions in order to examine the relative performance of all the models. The experimental results have demonstrated in the ability to predict the fatigue life and FCG rate. This study indicates that higher value of the initial crack length or aspect ratio leads to a lower number of cycles for FCG.
Moreover, an increase in the stress ratio and crack growth rates has been noticed. The ANFIS method provided the best fit to the FCG rate data while, the CANFIS results gave the highest fatigue life compared to the experimental results. The neuro-fuzzy prediction results showed that the under-predict methodology for the fatigue life was ± 0.019% for the CAL while between ± 0.018% and ± 0.027% for the VAL, which is conservative from reliability point of view.
The work carried as a part of Ph.D study entitled ‘Fatigue Crack Growth Prediction Using Experiment and Neuro- Fuzzy Method’. The assessment of the present results indicates that the prediction is highly reliable and conservative in comparison to other two fatigue models. Finally, the results show that most of the fatigue models have a different background and not capable of representing all factors affecting the fatigue life. The introduction of neuro-fuzzy method has solved those problems by avoiding construction of the explicit model equation and directly extracting the possible rules from the experimental data.
For further correspondence kindly contact :
Figure 1. The FCG worksheet based on Glyphwork software Figure 2. Prediction software interface with Fatigue Crack Growth rate utility
INNOVATE | Current Research
12
Potential of Banana Leaf as an Effective Lubricant for High Temperature Application
In Malaysia, the old generation had made full use of banana leaf not just for serving food but also for household use like smooth and effortless ironing. Similar to the famous non-wetting lotus effect, banana leaf also possesses a waxy adaxial surface in which the high contact angle of the water droplet is attributed to the microstructure of the leaf.
As such, various surfaces have been mimicking lotus leaf structure and rendering hydrophobic surfaces with high contact angles for various kind of applications. By understanding the combination of surface characterisation along with the adhesion and friction properties, one can allow improved hydrophobic surfaces.
Associate Professor Dr. Mariyam Jameelah bt Ghazali Department of Mechanical and Material Engineering Faculty of Engineering and Built Environment Universiti Kebangsaan Malaysia
43600 UKM Bangi, Selangor, MALAYSIA
Tel: 603-89216418 Email: mariyam@eng.ukm.my
For further correspondence kindly contact :Scanning electron micrographs of adaxial surface of banana leaves under 5 N at (a) 110 oC (b) 150 oC and (c) 200 oC
Though there are not much literature and research on the actual area involved, new research emerging from the current work is hoped to reinvigorate the use of banana leaves particularly in the field of tribology. In this work, several iron plates with different weight and temperatures were tested to determine the effectiveness of the banana leaves’
cleaning effects. In short, banana leaf could be one of the best candidates for anti-friction materials particularly for elevated temperature applications as it had been used to clean and smooth up dirty hot irons in the old days.
INNOVATE | Current Research
13
Mohd Yazmil Md Yatim
Department of Civil and Structural Engineering Faculty of Engineering and Built Environment Universiti Kebangsaan Malaysia
43600 UKM Bangi, Selangor, MALAYSIA Tel: 603-89118343 Email: mymy@eng.ukm.my
For further correspondence kindly contact :Development of Design Method for Thin-Webbed Composite Plate Girders with Partial Interaction
In a typical composite construction, steel plate girders and concrete slab are firmly connected together so that they can act compositely as a single unit, thus fully exploiting the advantages of the two different materials. Headed studs, the most widely used type of shear connector, are flexible in nature in the sense that they may deform to a certain extent depending on the magnitude of shear force transmitted through the connectors, thus inducing ‘partial’
condition in the composite interaction. Partial interaction permits longitudinal slip arising from strain discontinuity at the steel-concrete interface. Presence of slip gives rise to deflection due to additional curvature which in turn, lowers the flexural stiffness and strength of the composite members. However, partial interaction design becomes the only solution particularly when the area of top flange is insufficient to accommodate the number of headed studs required for full interaction. In addition, thin-walled members possess reserved strength in the plastic region. Combination of post-buckling effect and composite action results in complex behaviour of the girders. Analysis of such behaviour becomes complicated when the composite interaction is partial.
Recently, a series of experimental tests have been conducted on eight full-scale thin-webbed composite plate girder specimens. The main variables are the degree of composite interaction and concrete strength. The specimens were tested to failure under monotonic loading and extensive measurements of displacement and strain were carried out.
Significant variations of strength, behaviour and failure characteristic were obtained from the tests due to the effect of partial interaction. Based on these findings, a method may be proposed for design of partially connected composite plate girder, align with the provisions in the current design codes.
Figure 1. Behaviour of Composite Plate Girder at Failure Load
INNOVATE | Current Research
14 14
Professor Dr. Ahmad Kamal Ariffin Mohd Ihsan.
Department of Mechanical and Materials Engineering Faculty of Engineering and Built Environment
Universiti Kebangsaan Malaysia
43600 UKM Bangi, Selangor, MALAYSIA Tel: 603-89216500 Email: kamal@eng.ukm.my www.eng.ukm.my/cem
For further correspondence kindly contact :
Deterministic and Probabilistic Finite Element Analysis of Elastic Interacting Cracks
INNOVATE | Current Research
The analysis, prediction and prevention of failure in mechanical structures and components based on a single cracked body basis are no longer adequate to support the existing fracture mechanics approaches. For multiple cracked body basis due to occurrence of multiple cracks interaction in most of the manufactured and assembled parts that consist of multiple stress concentration features such as notches, corners, bends and holes. The available structural assessment guideline for cracked body, known as fitness-for-service (FFS) by the ASME boiler and pressure vessel
Figure 1: Computational framework code (Section XI, Articles IWA-3330), JSME fitness-for-
service code and BSI PD6493 still consider multiple cracks as one larger crack regards to certain rules and conditions.
To a certain extent of cracked body geometry, those codes are inaccurate. Due to this, engineers are unable to design the mechanical structures and components subjected to minimize the risk of failure, increase safety factor and life cycle. The available analytical solutions encountered many limitations with multiple assumptions and unable to represent the real physical and mechanics behavior of multiple cracks interaction. The advancement of computational fracture mechanics tools, able to solve at least two issues of multiple crack problems such as crack interaction limit and multiple cracks-single crack equivalencies. The analytical formulations are used to understand the fracture modes of brittle failure behavior promoted by crack interaction.
Assuming the inelastic deformation and other non-linear effects is relatively small compared to crack sizes and other geometrical body characteristic, the elastic crack interaction is used to address the crack interaction behavioral problems for multiple edge cracks based on the high stress gradient near the crack tip or stress intensity factor (SIF). To determine the SIF, the Linear Elastic Fracture Mechanics (LEFM) and Theory of Critical Distance (TCD) in continuum mechanics are used to develop two new algorithms based on displacement extrapolation method and strain energy release rate since
both methods are well established in single cracked body analysis. Modeling errors and numerical errors are justified to improve the accuracy and reliability. The new finite element discretization error estimation is formulated to improve the numerical convergence. The investigation has been successfully executed by a deterministic approach. The probabilistic finite element approach is used to formulate the uncertainty effect of elastic crack interaction in brittle failure, which still in progress. The developed finite element procedures are validated by an independent and dependent method. As a result, the two mentioned issues will be solved deterministically and stochastically with the inclusion of elastic crack interaction. Thus, provide more reliable analysis, prediction and prevention of failure for multiple cracked bodies.
15 INNOVATE | Current Research 15
Design of Dynamic Phase Frequency Detector with Charge Pump and Loop Filter for High Speed PLL
A phase locked loop (PLL) is one of the key modules in wide range of wired and wireless RF devices. The operation of PLL depends on the performance of its detector. Phase frequency detector (PFD) is the first element found in analog and digital PLL system. PFD plays an important role in detecting of any phase and/or frequency inputs difference.
As the multi-gigahertz system is going under rapid evolution, detectors need to be improved in achieving low power operations, accurate phase error detection and high circuit sensitivity. Besides, existing PFDs yield high gate delay and large layout area which affect overall PLL performance. The challenge is to overcome those shortcomings of the existing PFDs by using transistor sizing optimization method. The failure to operate at low supply voltage and also different current driving capabilities of transistors caused delay variations which lead to increase power dissipation (Koo 2005). Total power dissipation is proportional to the total width of the transistors in the signal path (Homayoun 2013). A method to measure dead zone is by looking into circuit sensitivity as an indicator of PFD capability in detecting phase differences (Chou 2004).
A dynamic PFD design by Kruti et al. 2011 has been utilized. CMOS dynamic PFD works efficiently to obtain multi- gigahertz PLL operations due to less number of transistors and consume low power (Mansuri 2002). The critical PFD parameters were identified. The basic rules for transistor sizing used as guideline are as follow (1) when a bigger channel length is chosen, optimized device width is important to produce the smallest output delay of dynamic PFD (2) keeping transistor’s Wp/Wn ratios equal to 2 at each path (3) keep the transistor length as minimum to keep lower input power supply which resulting to low threshold power, low delay and small design area (4) avoid additional transistors at delay path for a lower power consumption (5) if multiple MOS are in series make the ones furthest from the output larger (increase W/L) then the ones closest to the output which is called progressive sizing (Homayoun 2013). Preselecting the channel length is the most physically based and time-efficient way to identify the MOS transistor sizes in a circuit. The dead zone size is 2 ps which relatively very small due to the input rise time and input fall time used in this work is fixed at 1 ps. The optimized device sizes circuit design of this work successfully detect the PFD input phase difference regardless the input power supply voltage sizes.
The proposed dynamic PFD with charge pump and low pass filter simulation output waveforms and industrial standard dynamic PFD mask layout depicted in Figure 1(a)-(c). A simple dynamic phase frequency detector (PFD) with low input power supply of 1.2 V has been employed in 0.18 µm CMOS process to overcome those shortcomings of the existing PFDs. This design maintained the circuit stability when operate at low and very high input frequency.
The dead zone of undetected input frequencies phase difference of this design exhibits only 2 ps. The peak-to-peak spikes are as low as 76 mV which is insignificant to PLL overall performance. Moreover, by avoiding standard gate level design approach, the proposed dynamic PFD became highly compact occupying only 192 µm2 layout area which will further trim down the total cost of the PLL. Total power dissipation is 116 pW which is lower than other typical PFDs that gives the designers better options in developing high speed PLL devices.
For further Correspondence kindly contact :
Figure 1. Proposed optimized (a) testbench output waveforms (b) DFF layout (c) dynamic PFD layout
Professor Dr. Mamun Bin Ibne Reaz, Mohammad Marufuzzaman Department of Electrical, Electronic and Systems Engineering Faculty of Engineering and Built Environment
Universiti Kebangsaan Malaysia
43600 UKM Bangi, Selangor, MALAYSIA
Tel: 603-89118406, Email: mamun.reaz@gmail.com
16
16 INNOVATE | Current Research
Nanomaterial is a field that takes a materials science-based approach to nanotechnology. Nanotechnology is proposed to create various functional devices, materials, structures and systems in nanoscale dimension. In the current research works, nanomaterials were synthesized for solar cell fabrication, E. coli sensor and microstrip patch antenna. The research works on the solar cell fabrication included finding suitable metal/semiconductor metal oxide materials as anti-reflective coating and photoanode for dye-sensitized solar cell (DSSC). The main function of anti-reflective coating material is to absorb more sunlight and enhance the efficiency of the cell’s performance.
Moreover, dye-sensitized solar cell is a new approach in solar energy technologies to convert solar radiation into electricity. Simple fabrications processes, low cost and less hazardous to the environment are the main advantages of DSSC. A chemical bath deposition and sol-gel doctor blade techniques were implemented to synthesize and fabricate the DSSC’s photoanode. The basic optical, structural, chemical and photovoltaic properties of the metal oxides as a base material were analyzed to obtain high efficiency and low cost solar cell. For sensor research, the fabrication of lab-on-a-chip E. coli sensor based conducting polymer-metal and polymer-metal oxide nanocomposite thin films were conducted. The sensitivity of the sensor was tested using I-V measurement and impedance analysis.
Incorporation of metal or metal oxide nanoparticles into the conductive polymer matrix can increase the specific area of these materials and thus improve the sensitivity performance. In the microstrip patch antenna research, new dielectric materials were investigated to enhance the performance of the antenna and miniaturize the size of the antenna. The metal oxide-dielectric materials were proposed to provide different values of dielectric permittivity which can be applied as a patch material for global positioning system (GPS) and wireless communication system.
Nanomaterial for Functional Devices
Figure 1 (a) E. coli sensor , (b) Dye-sensitized solar cell and (c) Microstrip patch antenna
For further correspondence kindly contact :
Associate Professor Dr. Huda Abdullah
Department of Electrical, Electronic and Systems Engineering Faculty of Engineering and Built Environment
Universiti Kebangsaan Malaysia
43600 UKM Bangi, Selangor, MALAYSIA Tel: 603-89216310 Email: huda@eng.ukm.my
(a) (b) (c)
17 INNOVATE | Current Research 17
Thermal Cooling System for Electric Vehicle Li-Ion Battery
Most of the current electric vehicle power source is Li-ion battery which can produce high heat generation during charging and discharging process. As extreme temperature can affect performance, reliability, safety and lifespan of electric vehicles, thermal management of Li-ion battery system is critical to success of all electric vehicles. It is known that under normal power loads and ambient operating conditions, the temperature of most Li-ion cells can be easily controlled to within the range of 20-55 degree Celsius. However, during high power draw, high ambient temperature, or high power charging, local heat generation may be steeply increased. This may lead to propagation of excessive temperature throughout a module or pack and even thermal runaway. Therefore, a good thermal management system must be able to maintain the batteries’ temperature within a suitable range without being over- designed or limiting battery performance. The existing solutions are mainly based on the air cooling technology which has limitations in handling fast charging and discharging such that an individual failure cell could trigger thermal runaway throughout the battery package. On the other hand, liquid cooling systems provide more effective heat transfer and are more compact. Recently, mini/micro scale heat transfer is crucial in various thermal processes.
The higher heat transfer coefficient attainable via the use of small size channel coupled with the high surface-to- volume ratio enables mini/micro-channel heat sink to effectively dissipate a substantial amount of heat in liquid cooling system. To optimize the performance of a battery pack/module, the thermal energy management system which employs cold plate with liquid coolant is proposed as shown in Figure 1. An oblique mini-channel liquid cold plate was developed to cool electric vehicle batteries in cases of high heat dissipation. The designed liquid cold plate contained continuous straight fins with simple oblique cut at optimized angle and width. The segmentation of continuous fin into oblique sections leads to the re-initialization of boundary layers. Different arrangement of oblique fins is studied in order to enhance the rate of cooling.
The work carried out is part of the project entitled ‘Heat Transfer in Hybrid Single-Phase Jet Impingement Micro- Channel for High Heat Flux Cooling’ supported through the Fundamental Research Grant Scheme (FRGS) by the Ministry of Education, Malaysia.
Associate Professor Dr. Rozli Zulkifli Centre for Automotive Research (CAR) &
Department of Mechanical and Materials Engineering Faculty of Engineering and Built Environment Universiti Kebangsaan Malaysia
43600 UKM Bangi, Selangor, MALAYSIA Tel: 603-89216104 Email: rozli@eng.ukm.my
For further correspondence kindly contact :Figure 1. Schematic Diagram of Cold Plate Liquid Cooling System
18
18 INNOVATE | Current Research
Intelligent Cooperation Management among Base Stations with Renewable Power System for a Green Fifth Generation (5G) Wireless Network
Energy efficiency in cellular networks is a growing concern for cellular operators with regards to maintaining profitability and reducing environmental impact. Under low-traffic conditions, BSs are underutilised. Hence, switching BSs on/off in accordance with traffic pattern variations is considered to be an effective method of improving energy efficiency. Two main concerns of network operators when applying this technique are coverage issues and securing radio service for an entire area in response to the increased size of some cells to provide coverage for cell areas that are switched off.
In this research, we aim to develop a method to switch off/on the macro-BSs according to the traffic conditions, while guaranteeing service and coverage, as shown in Figure 1(i). The next phase of our research considers deployment and integration of future generation of wireless networks, i.e. 5G into the legacy LTE wireless networks, as illustrated in Figure 1(ii). It is expected to achieve balance between demand on energy efficiency and high data rate during peak traffic hours. Three key aspects will be investigated in this research: (i) energy yield analysis, (ii) economic analysis, and (iii) greenhouse gas emissions. This project is funded by the Centre for Collaborative Innovation UKM, under the Economic Transformation Plan (Ref. No: ETP-2013-072) grant scheme
For further correspondence kindly contact :
Ir. Dr. Rosdiadee Nordin
Department of Electrical, Electronic and Systems Engineering Faculty of Engineering and Built Environment
Universiti Kebangsaan Malaysia
43600 UKM Bangi, Selangor, MALAYSIA Tel: 603-89118402 Email: adee@eng.ukm.my
Figure 1. (i) LTE macro-BSs powered by a smart grid, and (ii) Deployment and integration of 5G into existing
( i ) (ii)
19 INNOVATE | Current Research 19
Wideband Planar Antenna Design Using Heuristic Optimization Algorithm
The rapid growth of wireless communication industry across the globe with the implementation of standards like WLAN, GSM and RFID has created a demand for multifunctional, low cost and compact antennas. The design of these complex antenna structures to achieve high gain, wide bandwidth and multiple operational frequency bands is almost impossible using traditional deterministic mathematical techniques as these structures often do not have any simple mathematical representation. Recently, heuristic algorithms, such as Particle Swarm Optimization (PSO) and Genetic Algorithm (GA), have been utilized, along commercial EM simulation software, to optimize complex structures.
As these algorithms often suffer from premature convergence and curse of dimensionality, a novel PSO variant, Ramped Convergence PSO (RCPSO), is proposed for optimization of compact microstrip antennas in this research.
A multi start approach is implemented to alleviate the premature convergence issue of PSO and the k-dimensional breakdown technique elegantly handles the curse of dimensionality in complex antenna structures. The purpose of the proposed algorithm is to design and optimize complex microstrip antenna structures for application in wireless communication industry, such as RFID, and aid the analysis of these structures. Heterogeneous boundary conditions are applied as compactness of the optimized antennas is crucial. The RCPSO algorithm is briefly described below:
1. Initialize the particles to random locations in the search space for first set of dimensions
2. Evaluate the fitness of each particle and update their personal bests along with the global best (similar to PSO) 3. Update the location of these particles based on the acceleration coefficients, personal bests and global best.
4. Implement heterogeneous boundary conditions on the swarm.
5. If predetermined number of generations is completed for current set of dimensions, re-initialize the swarm for next set of dimensions.
6. Go to step 2 and repeat.
The work carried out on RCPSO is part of the project entitled ‘A new hybrid optimization algothrim using adaptive weights based machine learning technique for wideband microstrip antenna’ and supported through the Fundamental research grant (FRGS). The proposed RCPSO algorithm takes into consideration i) reduction of fitness function evaluations and ii) quality of the optimized design iii) faster and accurate convergence. The introduction of RCPSO holds promise to improve optimization speed, provide adaptive design capabilities, offer the ability to identify key antenna structures and construct novel features and structures for microstrip antennas.
Professor Dr. Mohammad Tariqul Islam
Department of Electrical, Electronic and Systems Engineering Faculty of Engineering and Built Environment
Universiti Kebangsaan Malaysia
43600 UKM Bangi, Selangor, MALAYSIA
Tel: 603-89216857 Email: tariqul@ukm.edu.my
For further correspondence kindly contact :20 20
Prototype Implementation of Intelligent Smart Bin for Real Time Solid Waste Status Monitoring
At present, the volume of generation of municipal solid waste (MSW) is increasing rapidly with the increase of population, economic growth, industrial development, changes in consumption habit and life style of urban population. It becomes a great challenge to manage MSW to the authorities in charge for waste management. Due to the lack of proper management solution, a considerable amount of 85% of the total MSW management budget is exhausted on waste collection and transportation. A well-organized and proper way to monitor the status of solid waste bin in real time whereas confirming green environment and viable advance of the society. The proposed approach presents the design and development of a prototype of intelligent smart bin that can collect bin status parameters in real time while consuming minimal energy.
The work carried out is part of the project entitled ‘Prototype implementation of solid waste bin and truck monitoring and management system’ and supported by the Prototype Research Grant Scheme (PRGS). The proposed smart bin is composed of sensor node mounted on it for the data collection and transmission. A set of carefully chosen sensors are employed to collect bin information. Among the sensors, one group is mounted on the underside of the bin cover and the other is attached upside of a foil installed in the bottom of the bin. The first group consists of an accelerometer, a hall effect, an ultrasound, a temperature and a humidity sensor. The other group consists of the load cell sensor. The details setup of smart bin is shown in Figure 1. The sensors are integrated using a sensor board. Along with a ZigBee transceiver, all the sensing devices except the load cell are installed inside a box. The intelligent system takes into consideration i) real time response and ii) minimum energy consumption by developing an intelligent sensing algorithm. It is anticipated that the use of the real time bin status data from the designed system in route optimization will help to reduce solid waste collection costs and GhG emissions caused by waste collection vehicles.
For further correspondence kindly contact :
Professor Dr. M A Hannan
Department of Electrical, Electronic and Systems Engineering Faculty of Engineering and Built Environment
Universiti Kebangsaan Malaysia
43600 UKM Bangi, Selangor, MALAYSIA Tel: 603-89216328 Email: hannan@eng.ukm.my
Figure 1. Smart Bin Prototype
INNOVATE | Current Research
21 INNOVATE | Current Research 21
FPGA Based Human Activity Detection System
Intelligent home or smart home is a process to develop a managed environment for comfort, safely, healthcare and energy optimization. It possesses an intelligent unit to learn the environment and the inhabitant to control the house according to user requirements. Multiagent algorithms frequently used to solve the problem of ubiquitous environment especially in smart home. Ubiquitous environment (smart home) consists of numerous environmental parameters, which are subdivided into smaller problems to reduce complexities. To finalize the solution, smaller sub domains are integrated utilizing multiagent architecture. The Smart home research requires understanding of the human behavior and recognizing patterns of activities of daily living (ADL). Early projects in this area hardly try to understand psychosomatic nature of human. Those projects simply employed intelligence to the household appliance without considering psychological understanding. The algorithms were not very effective and the architectures were not strong enough to achieve desired progress.
Human activity always follows some patterns. Therefore, in order to predict the future tasks it is necessary to understand ADL. The intelligent home (IHome) project at the University of Massachusetts at Amherst developed a simulated environment designed with multiagent survivability simulator (MASS) and a java agent framework (JAF) as tools to evaluate agent behavior and their coordination. The project models agent interactions and task interactions to evaluate the tradeoff between robustness and efficiency. However, the project never builds a practical smart home to evaluate their model. Recently, Reaz et al. presented a multiagent system to develop an intelligent home controller.
The project presented a lower cost, portable solution to promote wide adaptation of smart home. The researchers only consider two agent spaces and did not discuss how the system will adopt more agents to facilitate larger environment. Besides, it is very important to present the algorithms in FPGA as the system can get a high-speed level as well as an exciting precision. In this research, an FPGA based intelligent home architecture is proposed using multiagent. This multiagent is designed based on tasks. The proposed algorithm solely considers appliance states, which is based on sequences characteristics and is able to identify different tasks of the resident accurately.
Detection of tasks is playing a critical role for forecasting the actions of resident as well as his behavior.
FPGA implementation of detection of individual tasks has been developed in Quartus II Altera environment. The hardware description language (HDL) used in this work is Verilog HDL. In order to evaluate the algorithm, a huge practical set of data is required. This research is used data from a practical smart home name as MavHome project.
In order to detect the tasks correctly MATLAB functions are used to filter the unwanted data. In Quartus II Altera environment, it is possible to simulate the design even with the gate level delays. ModelSim SE 6.3a is a widely used simulator for viewing and analyzing the simulation results of FPGA. Thus, it is used in this research to show the
Professor Dr. Mamun Bin Ibne Reaz, Mohammad Marufuzzaman Department of Electrical, Electronic and Systems Engineering Faculty of Engineering and Built Environment
Universiti Kebangsaan Malaysia
43600 UKM Bangi, Selangor, MALAYSIA
Tel: 603-89118406, Email: mamun.reaz@gmail.com
For further correspondence kindly contact :22 22
simulation output of detecting tasks. Figure 1 shows the Gate Level Simulation (GLS). The design runs in 20MHz clock frequency so each clock period is required only 50ns. In Figure 1 it is shown that if the reset signal is high i.e. “1”, no tasks are detected. It is also clearly shown that using 4 bits, 6 individual tasks can be representing successfully. A multiagent system is surely providing intelligence to the household appliances. MultiAgent algorithm is derived from AI, which is aimed to solve complex problems in ubiquitous environment especially smart home. Each of the agents is responsible for specific task, which is partially shared with a decision-making agent for problem solution. In future, the probabilistic values can be represented in fixed-point format as well as the predicted events. The expected outcome of this research is to have an efficient decision making controller.
FPGA Based Human Activity Detection System
For further correspondence kindly contact :
Professor Dr. Mamun Bin Ibne Reaz, Mohammad Marufuzzaman Department of Electrical, Electronic and Systems Engineering Faculty of Engineering and Built Environment
Universiti Kebangsaan Malaysia
43600 UKM Bangi, Selangor, MALAYSIA
Tel: 603-89118406, Email: mamun.reaz@gmail.com
Engineering & Built Environment
Ph.D Thesis 2013 Abstracts
25 25
Advanced Constitutive Modeling of Granular Aterials Incorporating Fabric Anisotropy Effects
Homayoun Shaverdi (P49894) by
Doctor of Philosophy in Civil and Structural Engineering (Ph.D)
A new formulation is presented to calculate the evolution of fabric under the rotation of principal stress axes of granular or sandy materials. The induced fabric anisotropy is attributed to parameters such as the shear to normal stress ratio, the internal mobilized friction angle, and the non-coaxiality between the stress tensor and the fabric anisotropy tensor. All these parameters are taken into account in the micro-level analysis. The second invariant of the fabric tensor is an essential component to account the effect of induced anisotropy. All the above mentioned parameters are incorporated in the equation proposed to calculate the second invariant of fabric anisotropy. It enables the constitutive modelers to determine the evolution of fabric and to include it in the constitutive equations.
A comparison with experimental tests demonstrates the validity of this formula for the evolution of anisotropy. The effect of fabric and its evolution is also included in the shear strength formulation of granular materials. The extended Mohr-Coulomb yield surface is developed to account for the inherent and induced anisotropy. Bounding surface plasticity theory is used to capture plastic deformation in the presence of inherent and induced anisotropies. In the existing constitutive models, to account for the inherent anisotropy and to determine the plastic modulus, the critical state line must be modified. However by applying these developed equations, there is no need to render the critical state line in the plane. In the simulations that was conducted in this study, the 32% reduction in the shear strength is well simulated by a one set of constants only. The great difference in the sand response due to different bedding angle and the deviation angle between the principal components of stress and fabric tensors is also well simulated by the model.
For further correspondence kindly contact:
Professor Dr. Mohd Raihan Taha (Main Supervisor) Department of Civil and Structural Engineering Faculty of Engineering and Built Environment Universiti Kebangsaan Malay sia
43600 UKM Bangi, Selangor, MALAYSIA
Tel: 603-89118373; Fax: 603-89118344; Email: drmrt@eng.ukm.my
INNOVATE | Ph.D Thesis 2013 Abstracts
26
Adopting Artificial Intelligences in Optimizing Reservoir Operation Policy
by
Md. Shabbir Hossain (P54849)
Doctor of Philosophy in Civil and Structural Engineering (Ph.D)
Reservoir release policy is a highly sensitive issue in water resources management as it directly controls the water uses in different aspect of a nation. Most of the cases consists nonlinear problem formulation, uncertain situations and complex operational conditions. The conventional methods, like dynamic programming (DP), linear programming (LP) and nonlinear programming (NLP) may fail to produce satisfactory optimum results in solving them due to their own limitations. DP is not suitable for a large dimensional problem and need discrete variable spaces, LP required modifying the objective function in a linear form and NLP is not convenient to use for a problem consisting a huge number of constraints and variable bounds which is very obvious in case of reservoir operation optimization.
In addition, other popular and well-established optimization algorithms like genetic algorithm (GA) and particle swarm optimization (PSO) have the formation limitations and disadvantages too for using in this type of optimization problem. GA has the complexity of using evolutionary operators and premature convergences problems and PSO may suffer from entrapping in local optima. This study has introduced a newly developed search algorithm named as artificial bee colony (ABC) optimization in solving nonlinear reservoir release optimization problems. Minimization of the water deficit for monthly operation has considered as the main objective of the study. While achieving the objective the study has proposed ABC in reservoir release optimization research as it free from the difficulties of previously used optimization techniques. The proposed ABC optimization has been compared with two leading optimization techniques in this field – GA and PSO. The study has investigated the performances of ABC, PSO and GA optimization models on two case studies – large and small reservoir, categorized in terms of reservoir physical and operational characteristics. In calibrating the release policy, three inflow situations (high, medium and low) have constructed from the historical data analysis for both reservoir. Simulation has done by using actual historical inflow to perform the risk analysis for each model. Different performance measuring indices (such as reliability, resiliency, vulnerability etc.) has computed and compared for this purpose. The results showed that for both types of reservoir (large and small) ABC outperformed in terms of reliability, resiliency and vulnerability. The ABC driven release policy has handled the small as well as large and complex reservoir system smoothly where GA and PSO has found not to be able to reach that much satisfaction level for large and complex reservoir. In case of small and simple reservoir the reliability measures are 74 %, 71.5 %, 70% and 52% respectively for ABC, PSO, real coded GA and binary GA model. For large and complex reservoir the reliability measures obtained from ABC, PSO, real coded GA and binary GA models are 98.14 %, 97.69, 91.66 % and 70.8 %. ABC release policy also has the capability to save water from wastages as the simulated results showed lowest number of oversupply period.
For further correspondence kindly contact:
Assoc. Professor Dr. Ahmed Hussein Kamel Ahmed Nasser Elshafie (Main Supervisor) Department of Civil and Structural Engineering
Faculty of Engineering and Built Environment Universiti Kebangsaan Malaysia
43600 UKM Bangi, Selangor, MALAYSIA
Tel: 603-89216220; Fax: 603-89118344; Email: elshafie@eng.ukm.my
INNOVATE | Ph.D Thesis 2013 Abstracts
27
Treatment of Produced Water Using Nanofiltration and Reverse Osmosis Membranes as a Possible Source of Potable Water
by
Salem Abdullah M Al Zahrani (P53246)
Doctor of Philosophy in Civil and Structural Engineering (Ph.D)
Produced water is one of the largest wastewater streams generated by the petroleum industry and totals more than 437 million barrels per day, of which 99% is disposed to the environment. Consequently concerns relating to water resource sustainability suggest exploring the potential reuse of treated produced water (viz., produced water treated to disposal standards) as a promising source of water. To that end, this research investigated the possibility of treating produced water using new and highly hydrophilic nanofiltration (NF) and reverse osmosis (RO) membranes for reuse as indirect potable water. The study comprised comprehensive characterization of membranes, evaluation of removal efficiency performance by assessing water quality and toxicity, and studies of fouling and cleaning. It was found that the NF membrane was super hydrophilic at 23° ± 0.90 and the RO membrane at 37° ± 0.49. Additionally, attenuated total reflectance Fourier-transform infrared spectra indicated that the membranes had a polyethersulfone structure. The NF membrane was characterized by a smooth surface, as given by a roughness average (Ra) value of 5.06 nm and a root-mean-squared roughness of 6.39 nm, whereas the Ra value of the RO membrane was 21.7 nm and the root-mean-squared roughness was 17.2 nm. The NF membrane exhibited higher permeability 7.3 (L. m–2. h–1) than the RO membrane 3.4 (L. m–2. h–1). The flux of filtration produced water reached 75 (L. m–2. h–1) through the NF membrane and 35 (L. m–2. h–1) through the RO membrane. The rejection of 12 single salts ranged 37%–96.6% for the RO membrane and 36%–97% for the NF membrane. A total of 213 parameters of water quality were investigated.
Water obtained with NF and RO membranes met 74 parameters listed in international standards of drinking water except those for molybdenum, ammonia and boron. Toxicity tests further revealed that the RO-treated produced water was free of toxicity and was therefore suitable for reuse. Conversely, the NF-treated produced water was found to be toxic at an effective concentration of 13.65% (EC50). The analysis of more than 138 hydrocarbon parameters not regulated internationally suggested that chemical additives have a high potential to contribute toxicity to produced water treated by an NF membrane. However, the NF membrane removed 48% of total organic carbon and its application showed potential as a pretreatment for the RO membrane and in other beneficial applications. The fouling studies showed that foulants in produced water are complex and originate from 975 mg/L total dissolved solids, 26.9 mg/L total organic carbon, 42,000 cfu/mL bacteria and 10 mg/L total suspended solids. Foulants had diameters ranging from 462.8 to 653 nm and their zeta potentials ranged from −21.0 ± 4.76 to −26.6 ± 8.22 mV.
Additionally, both biofouling and organic fouling reduced flux, and extracellular polysaccharides formed in biofilms.
Foulant size affected RO membrane more than NF membrane, while the latter were more affected by pressure variations, whereas high pH mitigated flux decline in both membranes. Chemical cleaning with NaOH, sodium dodecyl sulfate (SDS) and disodium ethylenediaminetetraacetate (Na2-EDTA) was highly effective in removing easily the fouling layer from membranes; efficiencies ranged from 81.33% to 142.9%. In conclusion, converting produced water from a pollution source into a new water resource was shown to be readily achievable, especially through the use of a combined-membrane system rather than a single-membrane system.
For further correspondence kindly contact:
Professor Dato’ Ir. Dr. Abdul Wahab Mohammad (Main Supervisor) Department of Chemical and Process Engineering
Faculty of Engineering and Built Environment Universiti Kebangsaan Malaysia
43600 UKM Bangi, Selangor, MALAYSIA
Tel: 603-89216410; Fax: 603-89118345; Email: wahabm@eng.ukm.my
INNOVATE | Ph.D Thesis 2013 Abstracts
28
Power Quality Monitor Placement and Voltage Sag Source Location Methods in Power Systems Using Multivariable Regression and
Genetic Algorithm
by
Asadollah Kazemi (P49868)
Doctor of Philosophy in Electric, Electronic and Systems Engineering (Ph.D)
Power quality (PQ) of delivered power to customers is becoming more important in recent years due to increased use of microprocessor, communication devices and tripping of electronic equipment with disastrous consequences in industrial plants. To improve power quality, it is necessary as a first step to identify the exact location of PQ disturbances so that any disputes among utility and customers can be resolved fairly. In this research, a new method for locating voltage sag source is proposed using multivariable regression (MVR) analysis, voltage deviation (VD) and standard deviation (SD) indices. Voltage sag is considered because it is the most common and problematic PQ disturbance. The accuracy of the proposed voltage sag source location method depends on accurate estimation of unmonitored bus voltages and the selection of power quality monitor (PQM) placement in a power system.
Accordingly, a new PQM placement method is developed for determining optimal number and placement of PQMs using voltage sag data, Cp statistic and genetic algorithm and it is called as the GACp method. The objective function of the GACp method is to minimize the Cp index. The advantage of the proposed method is that it is not sensitive to set voltage threshold value of PQMs. The voltage sag data is obtained by simulating several test systems, namely, the IEEE 9, 30 and 69 bus test systems using the DIGSILENT software. After determining optimum number and placement of PQMs, regression coefficients are calculated by using MVR analysis and considering voltages at the PQM locations. Then, relationship between all of the unmonitored buses and PQMs are estimated. Finally, the location of voltage sag source is identified by the VD and SD indices. The accuracy of the GACp method for PQM placement is compared with the modified binary particle swarm optimization method by carrying out simulation tests on the IEEE 30 bus test system. The obtained results show that the GACp method provides a good PQM coverage even for very shallow voltage sags in the system. In addition, to validate the proposed MVR voltage sag source location method, the IEEE 30 bus test system is used and compared with the sag location method based on the branch current deviation (BCD). The test results show that the MVR method can identify the exact location of voltage sag, but the BCD method can only track the path or a confined area where voltage sag is most likely located. Accordingly, the proposed methods would be useful for assisting power engineers in finding exact location of voltage sag so that necessary actions can be taken to solve PQ problems.
For further correspondence kindly contact:
Professor Dr. Azah Mohamed (Main Supervisor)
Department of Electrical, Electronic and Systems Engineering Faculty of Engineering and Built Environment
Universiti Kebangsaan Malaysia 43600 UKM Bangi, Selangor, Malaysia
Tel: 603-89216323; Fax: 603-89118359; Email: azah@eng.ukm.my
INNOVATE | Ph.D Thesis 2013 Abstracts
29
An Efficient Slow-Start and Congestion Avoidance Mechanisms for Transmission Control Protocol
by
Ghassan Akram Abed (P53024)
Doctor of Philosophy in Electric, Electronic and Systems Engineering (Ph.D)
Transmission Control Protocol (TCP) is the dominant protocol in modern communication networks, within which the issues of reliability, data flow and congestion control must be handled efficiently. TCP provides a communication service at an intermediate level between an application program and the Internet Protocol (IP). With the evolution of Third Generation (3G) toward 4G wireless network, such as Long Term Evolution-Advanced (LTE-Advanced) standard, the use of TCP becoming more important for reliable end-to-end data delivery. As TCP was initially designed for wired networks, the LTE-Advanced network performance will degrade due to signal impairment and heavily congested links. Consequently, the traditional TCP versions such as Tahoe, Reno, Newreno, Sack, Fack, and Vegas, could not fulfil LTE-Advanced requirements to support huge number of user equipment’s that need wider channel bandwidth, low latency, and the enormous data stream transferred over this network. The objective of this thesis is to develop an efficient congestion control technique during slow-start and congestion avoidance phases for new TCP agent over LTE-Advanced network. Instead of using exponential increments as in traditional TCP variants, the proposed slow-start algorithm duplicates the congestion window (cwnd) and then approximates cwnd size using a quadratic interpolation approach until reaching the slow-start threshold (ssthresh). An improved Additive Increase Multiple Decrease (AIMD) algorithm control cwnd increment to avoid the congestion by deriving a new relationship between AIMD parameters. Moreover, multiple TCP technique achieves virtual multi-flows estimated dynamically according to network congestion conditions over the same connection. The proposed congestion control mechanism is integrated within a new high speed TCP agent called TCP Rapid using Network Simulator 2 (NS-2) and is verified over a traffic model of an LTE-Advanced network. Performance comparison in term of throughput, packet loss, queue size and goodput, as well and the behaviour of the congestion window proved that the proposed mechanism outperform the standard TCP variants. The throughput of TCP Rapid increased by 3 folds compared to Tahoe and doubled with Newreno while the goodput is 45% to 57% better than Reno and maintaining minimum packet loss and adequate queue size. Finally, it can be concluded that the proposed TCP is substantially efficient for wide-bandwidth low-latency networks such as LTE-Advanced, thus achieving higher throughput and a lower packet loss level.
For further correspondence kindly contact:
Professor Dr. Mahamod Ismail (Main Supervisor)
Department of Electrical, Electronic and Systems Engineering Faculty of Engineering and Built Environment
Universiti Kebangsaan Malaysia 43600 UKM Bangi, Selangor, Malaysia
Tel: 603-89118392; Fax: 603-89118359; Email: mahamod@eng.ukm.my
INNOVATE | Ph.D Thesis 2013 Abstracts
30
A Framework to Assist In Selection and Prioritization of Operational Key Performance Indicators in Engineering Maintenance Organization
by
Ibrahim Yousef Arafat Sokar (P42352)
Doctor of Philosophy in Electric, Electronic and Systems Engineering (Ph.D)
The lack of effective deployment methods for managers to identify the relevant key performance indicators (KPIs) has caused numerous management problems in the planning and development stage of introducing a performance measurement system (PMS). Therefore, a structural approach in the selection of the most relevant operational KPIs needs to be identified and used in the PMS to support the organizational objectives. Factors influencing the performance of operations and the production path process should be identified and measured.
Developing a framework for incorporating the interrelationship between various factors influencing performance, interrelationships between influencing factors and KPIs, and interaction with feedback and tradeoff among KPIs is necessary. Unfortunately, no practical method for selecting the most relevant operational KPIs in engineering maintenance organization exists. The aim of this research is to develop a practical framework in the selection of relevant KPIs by considering the interrelationships, feedback, and tradeoff. The research was carried out in three stages. Stage I describes the Quality Function Deployment (QFD) model to prioritize the influencing factors that may affect organizational performance, and the initial selection of KPIs based on the degree of impact on the organization’s objectives. Furthermore, based on cause and effect, the “whats” and “hows” in QFD are translated into a measurement matrix that contributes to organizational success. In Stage II, the KPI selection is considered as a multi-criteria decision-making problem. This problem has been modeled according to the Analytic Network Process (ANP) method because it considers criteria and subcriteria relations and interrelations in selecting the KPIs. The opinions of many experts are obtained while building the ANP model for KPIs. When the ANP and artificial neural network (ANN) models are integrated, a KPI set selection can be easily made based on the opinions of one single expert. In this case, the geometric mean calculation of answers obtained from many experts is unnecessary.
Additionally, the effect of subjective Judgment of one single decision maker is avoided. An ANN model has been designed and trained using ANP results to calculate KPI priority. Based on the results, the ANN model can be used in the selection of KPIs for another new decision. From Stages I and II, 13 KPIs were selected and applied in a case study company to measure its performance. From the 27 factors influencing the organization, the most influencing are material availability, job ending on time, crew absence, staff utilization, and productivity time. The higher priority KPIs are productivity, throughput, utilization, and efficiency. Stage III presents a framework using the Fuzzy Cognitive Maps (FCM) method to quantitatively analyze the tradeoff among the 13 KPI targets resulting from Stages I and II. An FCM model can help the top management and decision makers to identify the behavioral improvements of one KPI target on the other KPIs and in the PMS as a whole. A different scenario was conducted as a case study in one of the engineering organizations to analyze the interconnections between the KPIs and the impact of changing the targets.
The result showed that productivity KPI has the highest tradeoff with backlog KPI among all the others. The FCM has been effectively used in decision making and the simulation of complex PMS, and can help the management in adjusting the KPI targets.
For further correspondence kindly contact:
Assoc. Professor Dr. Mohd Yusoff Jamaluddin (Main Supervisor) Department of Electrical, Electronic and Systems Engineering Faculty of Engineering and Built Environment
Universiti Kebangsaan Malaysia 43600 UKM Bangi, Selangor, Malaysia
Tel: 603-89118393; Fax: 603-89118359; Email: dryusoff2020@yahoo.com
INNOVATE | Ph.D Thesis 2013 Abstracts
31
A Real Time Automated Vehicle Incident Detection Using High Accuracy Global Positioning System
by
Ishak Mohamad (P42465)
Doctor of Philosophy in Electric, Electronic and Systems Engineering (Ph.D)
The Malaysian Institute of Road Safety Research (MIROS) reported that in 2009, there were 397,330 road accidents in Malaysia and half of them were occurred on municipal roads. In order to avoid major traffic congestion and other consequent incidents, an efficient incident detection system is therefore essential. Although various incident detection systems have been deployed such as laser, multi-sensors, closed-loop detector and Global Positioning System (GPS), the systems are still not fully automated, require longer delay for detection, have high operational cost and high detection error. Moreover, GPS is only used for navigation or positioning and not fully utilized for incident detection. The objective of this research is to develop a new and efficient real time automated incident detection (AID) system with short time to detect, no false alarm and higher detection rate. The research involved three phases;
data collection, AID algorithm development and verifications through simulation and drive testing. During static data collection, GPS signal availability and reliability at four locations in the study area were checked. In the mobile data collection, the GPS probe vehicle gathered various information such as vehicle position, direction, and speed while driving along municipal trunk roads in Bandar Baru Bangi, Selangor. The data were transmitted via ultra-high frequency (UHF) radio band to a monitoring station (MS). The AID algorithm detected abnormal measures based on user input threshold of the vehicle’s maximum speed, maximum acceleration and maximum deceleration from mobile data collection. The algorithm was also verified through simulation and the result could be viewed from the AID Interface (AIDEG) using random data for the normal and abnormal condition extracted from recorded mobile data. The system performance indicator includes time to detect (TTD), false alarm rate (FAR) and detection rate (DR). Finally, real time data from five GPS equipped vehicles along the pilot route were sent to the MS to test the performance of the AID system. The static test shows GPS availability and reliability rate of 99.9% and up to 99.7%
respectively, thus the study area is appropriate for AID application. Algorithm parameters obtained from mobile data are 33.33 m/s for the maximum speed, 3.24 m/s2 for the maximum acceleration and 4.63 m/s2 for the maximum deceleration. The result for real time monitoring shows excellent value of performance indicators; 3.3 µs to 3000099 µs TTD, zero FAR and 81% and above DR. However, the performance of the system depends on the availability of GPS signal and data transmission from the vehicles to the MS.
For further correspondence kindly contact:
Professor Dr. Mohd Alauddin Mohd Ali (Main Supervisor) Department of Electrical, Electronic and Systems Engineering Faculty of Engineering and Built Environment
Universiti Kebangsaan Malaysia 43600 UKM Bangi, Selangor, Malaysia
Tel: 603-89216302; Fax: 603-89118359; Email: mama@eng.ukm.my
INNOVATE | Ph.D Thesis 2013 Abstracts
32
Development of ZnxCd1-xS Window Material for High Efficiency Ultra- Thin Cadmium Telluride (CdTe) Solar Cell
by
Md Sharafat Hossain (P558
