proceedings ofthe fabrication of different types of 3d-structured surfaces (i.e. aspherical,...

Proceedings of

ISE Research Students' Conference 2013

research makes a humble heart

Department of Industrial and Systems Engineering

The Hong Kong Polytechnic University

Published by the Department of Industrial and Systems Engineering, The Hong Kong Polytechnic

University.

Copyright © 2013 Department of Industrial and Systems Engineering, The Hong Kong

Polytechnic University. All rights reserved. No portion of this publication maybe reproduced, by

any process or technique, without the express written consent of the copyright owner.

Preface

Welcome to the ISE Research Students' Conference 2013. I feel

honoured to write a Preface for these valuable conference

proceedings. The great success of the event would not have been

possible without the generous support and encouragement from

our department, the financial support from the VP(RD) via the

Faculty of Engineering, funding support from PolyU - Hotel

ICON Co-sponsorship Scheme and most importantly, our fellow

research students' participation and dedication. This year, the

conference brings over 50 research students together with

academics from 4 main disciplines covering Advanced Manufacturing Technology (AMT),

Knowledge and Technology Management (KTM), Logistic Engineering (LE) and Product and

Process Design (PPD). This event encourages the exchange of ideas among our research

students and forms the heart of the conference.

For 2013, the event carries a special theme - "research makes a humble heart''. To me,

starting my research study brings me to a different level of humbleness. We are working so

hard everyday digging into our research problems, making hypothesis and trying to find

solutions. It becomes inevitable that sometimes we may stumble and we feel we are not able

to go further. I come to identify truly with the saying that "the more one knows, the more one

feels insignificant". My research life here is a wonderful experience of learning to be humble.

I would like to express my sincere appreciation to all the organizing committee members for

their support, effort and enthusiasm. I would also like to thank the advisory committee,

especially Prof. T.M. Yue, Miss Cammy Chiu, Miss Iris Ko and Miss Meina Cheng for their

constant guidance and advice during the whole conference preparation. Finally, I would also

take this opportunity to thank all of you for your participation in the conference. I hope you

all enjoy the presentations and have a memorable time.

Mr Man Tik Dickson, Choy

Chairperson

Organizing Committee

ISE Research Students' Conference 2013

Organizing Committee

Chairman

Choy Man Tik, Dickson (Mr)

Secretary

Xin Ying, Anita (Miss)

Session Chairman / Chairlady

Chong Po Fat, John (Mr)

Ma Hoi Lam, Helen (Miss)

Muhammad Aamir Khan (Mr)

Xiao Gaobo (Mr)

Editorial Unit

Guo Daoqin (Mr)

Mak Chung Hong, Markson (Mr)

Ridvan Aydin (Mr)

Advisory Committee

T.M.Yue (Prof.)

Cammy Chiu (Miss)

Meina Cheng (Miss)

Administrative and Technical Support

Iris Ko (Miss)

Samuel Ngan (Mr)

Table of Contents

Advanced Manufacturing Technology (AMT)

Modelling and Optimization of Surface Generation in the Computer

Controlled Ultra-precision Polishing (CCUP) of Three Dimensional

Structured Surfaces

Cao Zhong Chen, Charles 1

Synergistical Enhancement of the Strength and Wear Resistance of

Titanium Alloys by TiN Surface Network Reinforcement

Chan On Ki 2

Design and Fabrication of Intelligent Scaffolds for Bone Tissue

Engineering

Chen Ling 3

Deformation Behavior of Bulk Metallic Glasses under Complex Stress

States

Chen Shunhua 4

Modeling and Optimization of Biomimetic Structures for Self-cleaning

in Ultra-precision Machining

Cheng Cheung Tong, Thomas 5

Microwave Processing of Titanium-Based Implants for Bone Tissue

Engineering

Choy Man Tik, Dickson 6

Wear Analysis of Diamond-coated Tools in Precision Machining of

Difficult-to-cut Materials

Fung Kai Yin 7

Magneto-caloric Effect of Fe-based Metallic Glasses at Room

Temperature

Guo Daoqin 8

Chemophysical Analysis of Composite Materials Due to Erosion

Damage

Huang, Wenfei 9

The Research of Ultra-Precision Light Field Measuring Microscope

Li Da, Adam 10

An Investigation of Processing Technology for Ultra-precision Roller

Embossing of Optical Micro-structured Surfaces

Lui Sing Yuen, Jasper 11

Fabrication of Bioactive Titanium Oxide Coating on Nickel-Titanium

Using Plasma Electrolytic Oxidation

Siu Hin Ting 12

A Multiscale Modelling Approach for Predicting the Critical

Undeformed Chip Thickness for Ductile Regime Dutting of Brittle

Materials in Ultra-Precision Machining

Xiao Gaobo 13

Re-manufacturing of Aeronautical Components by Additive

Technology

Xin Ying, Anita 14

Modeling and Proactive Resilient Self-adaption of the Effect of Tool

Wear in Ultra-precision Raster Milling

Zhang Guo Qing, Peter 15

Knowledge & Technology Management (KTM)

A Study on the Relationship between Knowledge Management and

Organizational Performance in the Manufacturing Industry

Khan, Muhammad Aamir 17

Computational Organizational Narrative Generation (CONG) for

Decision Support Learning

Yeung Chui Ling, Charlie 18

Assessment of Uncertainty in Quality of Knowledge in Research

Publications

Farzad Sabetzadeh 19

A Design Based Research to Conduct Knowledge Audit for

Unstructured Business Processes

Yip Yuen Tung 20

Theoretical and Experimental Investigation of Nano-surface

Generation in Ultra-precision Freeform Polishing: Process Modelling

and Optimization

Ho Lai Ting, Lesley 21

Intelligent Reporting of Intellectual capital for Value Creation in

Knowledge-intensive Organizations

Cai Linlin 22

A Study of Process Optimization and Nano-surface Generation in

Ultra-precision Machining of Precision Rollers for Advanced Optics

Manufacturing

Mak Chung Hong, Markson 23

A Computational Organizational Modeling and Simulation (COMS) for

Technology Assessment and Forecasting of Technology Intensive and

Innovative Enterprises

Cheng Mei Na, Meina 24

Post-adoption Behavior for Personal Learning Environment &

Network

Tsui Lai Na Miriam 25

Development of an Knowledge-based System for Managing

Competitiveness and Market Leadership of Project and

Process-Oriented Organizations

Rozhkov, Mikhail 26

Intellectual Capital and Value Creation – A Paradigm Shift?

Mariza Tsakalerou 27

Modeling and Ultra-precision Machining Micro-functional Structures

for Heat Exchanger

Wang Haitao 28

Development of an Intellectual Capital Driven Knowledge Audit

Methodology with Application

Gu Jie, Jessica 29

Logistics Engineering (LE)

An Investigation on Closed-loop Supply Chain using Priority Based

Genetic Algorithm Approach

Chen Yongtong, Cathy 31

Integrated Planning of Berth Allocation and Quay Crane Scheduling

Problems

Ma Hoi Lam 32

Integrating Production Scheduling and Mold Maintenance Planning:

An Genetic Algorithm Approach

Wong Chun Sing, Sing 33

Storage Allocation and Yard Trucks Scheduling in Container

Terminals using a Genetic Algorithm Approach

Wang Zhengxu 34

Production and scheduling in Supply Chain Management with

Uncertainty

Li Nan 35

An Integrated Green Supply Chain Framework for Sustainable

Industrial Development

Zhang Shuzhu 36

An Approach for the Perishable Product Logistics Based on Real-time

Monitoring with Radio Frequency Identification (RFID)

Wang Lixing 37

Design and Optimization of RFID-enabled Wireless Sensor Network

(WSN) Monitoring System for Biological and Pharmaceutical Products

Supply Chain

Ng Chun Kit, Felix 38

A Decision Support System for Managing Performance of Logistics

Service Providers in Cross-Border Operations

Lam Hoi Yan, Cathy 39

A RFID-based Resource Allocation System for Garment

Manufacturing

Lee Kar Hang Carmen 40

A Data Mining and Optimization-based Real-time Mobile Intelligent

Routing System for City Logistics

Lin Canhong, Jason 41

Enterprise Supply Chain Planning under Uncertainties

Liu Hongguang 42

Product & Process Design (PPD)

An Integrated Marketing and Engineering Approach to Product Line

Design

Ridvan Aydin 45

Prognostics of Chromaticity State for Phosphor-converted White Light

Emitting Diodes Using an Unscented Kalman Filter Approach

Fan Jiajie, Jay 46

A Flexible Capacitive Micromachined Ultrasonic Transducer (CMUT)

Array

Chong Po Fat, John 47

Modelling of Customer Satisfaction and Determination of

Specifications for Product Design Using Computational Intelligence

Techniques

Jiang Huimin 48

In-Process Visualisation for Deformation Diagnosis in Hydroforming

Kot Wai Kei Ricky 49

Failure Analysis of Titanium Tailor-welded Blanks under Multi-stage

Forming Process

Lai Chi Ping 50

Design of a New Molding Process for Making Seamless Hollow Plastic

Parts

Ng Wai On 51

A New Production Model to Compensate Forecast Error and Customer

Loss in Waiting

Qian Chen 52

A Flexible 2D Piezoresistive Shear and Normal Force Sensor Array

for Pressure Mapping Applications

Shi XiaoMei, Sissi 53

A System Monitoring Model by Examining Entity Dynamics

Wang Lei 54

A Novel Metaheuristic Model with Distributed Pattern Learning

Xue Fan 55

Advanced Manufacturing Technology

1

Project Title: Modelling and Optimization of Surface Generation in the

Computer Controlled Ultra-precision Polishing (CCUP) of Three

Dimensional Structured Surfaces

Name of Student: Cao Zhong Chen, Charles

Degree: PhD

Chief Supervisor: Prof. Benny Cheung

Co-supervisor(s): N/A

Contact Email: [email protected]

Office: DE406

Country of Origin: China

Nowadays, 3D-structured surfaces are widely used in various applications such as self

adhesive sensors, compound lenses for holography in phonics, etc. The fabrication of different

types of 3D-structured surfaces (i.e. aspherical, free-form) have always been a challenge to

the optical fabrication industry. Although there are some research studies on different types of

3D-structured surfaces, the research work on the fabrication of these surfaces is far from

complete. Fluid Jet Polishing (FJP) is a promising ultra-precision machining process that uses

an inclined adjustable nozzle to guide a premixed slurry to the workpiece at appropriate

speed. It is noteworthy that FJP can obtain a stable and controllable Gaussian-profile removal

function with no tool wear and polishing position sensitivity. Therefore FJP shows a potential

applied value in generation of 3D-structured artifacts and patterns for 3D-structured surfaces.

This is particularly true for machining difficult-to-machine and ferrous materials which are

not amenable by using other ultra-precision machining technologies such as single-point

diamond turning and ultra-precision raster milling.

For the FJP system, the material removal mechanisms, influenced by several operation

parameters (i.e. the slurry concentration, the particle size, the particle type, the slurry pressure,

the machining time, the impact angle, the standoff distance and the kind of workpiece

material), can be due to the collision and shearing actions between the abrasive particles and

the workpiece. To date, no one model has been developed that could contain all of these

operation parameters to accurately predict the material removal rate. Research work on

polishing mechanics, especially nano-mechanics, is far from complete, and little attention has

been focused on the generation of 3D structured artifacts and patterns for 3D-structured

surfaces by using FJP.

As for further research, the proposed study aims to establish a model-based simulation system

for the modelling and optimization of the surface generation in CCUP of 3D-structured

surfaces. The research will be developed by using a combination of standard laboratory based

experiments, a theoretical model and Computation Fluid Dynamics (CFD) Simulations to

predict material removal rate. The surface topography simulation model will take into account

the cutting mechanics, material removal mechanics at the nano-scale, material related factors,

kinematics and the dynamic characteristics of the polishing processes. The optimization

model allows the determination of an optimum polishing strategy.

Keywords: 3D-structured surfaces, CCUP, surface generation, CFD

2

Project Title: Synergistical Enhancement of the Strength and Wear Resistance

of Titanium Alloys by TiN Surface Network Reinforcement

Name of Student: Chan On Ki

Degree: MPhil

Chief Supervisor: Prof. H.C. Man



Office: EF403

Country of Origin: China (HKSAR)

High strength to weight ratio, excellent corrosion resistance and biocompatibility allow

titanium and its alloys to be used in various applications including aerospace areas, marine

areas and medical equipment. Other advantages include their good erosion resistance,

formability and high temperature capabilities. Titanium and its alloys are now widely used in

biomedical devices such as artificial bones, artificial joints, dental implants and artificial

blood vessels.

However, many researchers believed that the application of titanium and its alloys is limited

in tribological conditions. The primary drawback of titanium based alloys is its poor

resistance to sliding wear or fretting wear. Titanium and its alloys form a passive oxide layer

after contacting air spontaneously. However, the surface passive film is thin, and research

showed that the oxide layer is only 12-16 Å after immediate interaction with air, the thickness

only increases to 50Å after certain period of growth. According to ASM handbook concerning

titanium and its alloys, the oxide film will not reform after damage. Therefore, the protection

and surface performance of titanium and its alloys are not enough in using only their native

oxide film in applications. The relatively low hardness and poor wear resistance prohibit their

application in triboligical conditions.

This research project investigates the feasibility of applying laser gas nitriding techniques to

improve the wear resistance and strength of titanium alloys. In the present work, a continuous

wave fibre laser was used for the laser gas nitriding process under a pure nitrogen

environment on a substrate of commercial titanium alloy plate. A reinforcement network grid

of titanium nitride (TiN) tracks with various densities produced by laser gas nitriding is

fabricated on the alloy surface. The TiN network grid is metallurgically integrated into the

TiN alloy matrix at the surface. It is postulated that by varying the TiN grid density the

corrosion resistance could be improved to a level similar to the material with the whole

surface treated by laser gas nitriding. Compared with laser gas nitriding of the whole surface

of the material, it is believed that surface grid reinforced MMC would have a lower

production cost and processing time.

Keywords: Titanium nitride, Laser gas nitriding, NiTi, Biomaterials

mailto:[email protected]

3

Project Title: Design and Fabrication of Intelligent Scaffolds for Bone Tissue

Engineering

Name of Student: Chen Ling

Degree: PhD

Chief Supervisor: Prof. C.Y. Tang

Co-supervisor(s): Dr C.P. Tsui

Dr D. Z. Chen (Shenzhen University)


Office: EF403


Multifunctional scaffolds which have controllable biodegradability, drug delivery function,

and shape memory effect have been attracting great research interest in the field of bone

tissue engineering. These kinds of scaffolds can not only act as a physical support but also

provide the benefits of high therapeutic efficacy to accelerate bone regeneration.

Biodegradable polymer, poly-D-L-lactide (PDLLA), has been widely used in the

development of exogenous matrices suitable for facilitating tissue regeneration due to its

biocompatibility and biodegradability. Nevertheless, scaffolds derived from un-modified

polymers lack osteoconductivity and mechanical strength. Hybridization of polymer and

hydroxyapatite (HAp) was used to solve this problem.

In the present study, a novel technique combining polymer coagulation, cold compression

moulding, salt particulate leaching and a drug coating method was developed to fabricate

poly(ethylene glycol)/dexamethasone (PEG/Dex) coated porous PDLLA/nano-HAp

scaffolds. These scaffolds possess homogenous pore networks with high porosity and

controllable pore size. The mechanical properties of the scaffolds filled with nano-HAp were

dramatically improved. The surface hydrophilicity of the scaffold was significantly improved

by poly(ethylene glycol)/dexamethasone coating and nano-HAp addition. In vitro evaluation

of the mechanical properties, bioactivity, biodegradability and drug release behaviors of

PEG/Dex coated PDLLA/nano-HAp scaffolds has shown that the drug release behavior of

the scaffolds could be adjusted by varying the porosity level and nano-HAp incorporation

amount. Apatite detected on the scaffolds after exposure to a simulated body fluid showed

improvement in bioactivity and the apatite formation ability through the addition of the

nano-HAp content in the composites. Nano-HAp incorporation and apatite formation made a

positive impact on the mechanical properties of the scaffolds; however, plasticization and

degradation of PDLLA had a negative impact. The pH-compensation effect of the composite

scaffolds can reduce the risk of chronic inflammation complications. Scaffolds with good

shape recovery rate, thermomechanical properties, and degradability were investigated in this

study. Some other characteristics such as cytotoxicity will be investigated in a future study.

The fabrication method in this study can produce scaffolds with controllable structure,

appropriate mechanical properties, low degradation rates and good shape memory effect for

cancellous bone repair applications.

Keywords: Bone Scaffold, Biodegradable, Shape Memory Composite, Drug Delivery

4

Project Title: Deformation Behavior of Bulk Metallic Flasses under Complex

Stress States

Name of Student: Chen Shunhua

Degree: PhD

Chief Supervisor: Prof. K.C. Chan

Co-supervisor(s): Prof. L. Xia (Shanghai University)


Office: DE404


Due to the amorphous atomic structure, bulk metallic glasses (BMGs) exhibit unique

properties, such as high corrosion resistance, excellent mechanical and physical properties

and attractive processing potentials. The well known Achilles' heel of BMGs that impedes

their applications as structural materials is the limited overall plasticity at room temperature,

especially under tensile loading. With extensive effort, many techniques have been

developed to enhance the compressive plastic deformation behavior of BMGs, such as

composition optimization, dual phase microstructure tailoring and geometry confinement.

However, most of the research focuses on the deformation behavior under uniformly

distributed stresses. However, in practical applications for structural materials, they seldom

deform under uniform stress states, but mostly deform under complex stress states. For

instance, some recent research efforts have been spent on examining the deformation

behaviour of BMG structures, such as BMG foams and honeycombs. The results reveal that

those BMG structures exhibit ultra-large plasticity under compression tests, demonstrating

significant potential for engineering applications. Nevertheless, the deformation behavior of

BMG structures with complex stress states under tensile loading has less been studied and is

worthy of further investigations.

In this project, the prime aim is to shed light on the deformation behavior of BMGs under

complex stress states and give insights into the deformation mechanisms of BMG structures

under tensile loading. To achieve this, some BMG structural elements, such as Z-shaped and

curved BMG specimens, were firstly investigated to understand the plastic deformation

mechanisms of BMGs in complex stress states. The results have demonstrated, for the first

time, that with stress gradients, some BMG structural elements are able to exhibit plastic

elongation under tensile loading. Moreover, in complex stress states, the curved structural

element has demonstrated a three-stage deformation mechanism with large plastic

deformation. The findings have provided fundamental knowledge for further understanding

the deformation mechanisms of BMG structures.

In future work, combining experimental observation and FEM analysis, a theoretical study of

the deformation mechanisms of BMGs under complex stress states will be conducted. With

this basic knowledge, complicated BMG structures will be produced and examined. FEM

analysis will be used to predict and understand the deformation behavior of the BMG

structures. In addition, the deformation mechanisms of BMG structures will be investigated

for the potential structural applications of BMGs.

Keywords: bulk metallic glasses (BMGs), deformation behavior, complex stress states, shear bands

5

Project Title: Modeling and Optimization of Biomimetic Structures for

Self-cleaning in Ultra-precision Machining

Name of Student: Cheng Cheung Tong, Thomas

Degree: PhD

Chief Supervisor: Dr Sandy To

Co-supervisor(s): Prof. Benny Cheung

Prof. W.B. Lee


Office: DE404


Biomimetic structures such as lotus leaf and butterfly wing-like structures with self-cleaning

properties are becoming popular due to their extensive potential applications, such as for

vehicle windshields and exterior paint for buildings. In nature, leaf surfaces of numerous

plants are superhydrophobic and exhibit extreme water-repellence. Such repellence can be

applied to provide a surface with certain self-cleaning properties that could prevent dirt from

sticking to such surfaces. Most of the recent research on these areas concentrated on reporting

manufacturing methods, with detailed results of their contact angle measurement on the

sample surfaces without consideration the optical performance. Other than that, limited work

has been done on explaining the physics as to why micro-patterns could transform a

superhydrophilic surface to becoming superhydrophobic. This shows that extensive research

on superhydrophobicity has been collecting data for the sake of manufacturing rather than to

grasp a thorough understanding of the mechanism of self-cleaning taking place behind the

scenes.

This research aims to study the influence of surface geometries in ultra-precision machined

micro patterns on self-cleaning properties and to enhance understanding on how the

transformation could have taken place. An optimization model of biomimetic structures for

self-cleaning in ultra-precision machining will be developed to identify the optimal set of

parameters for the surface geometries of self-cleaning surfaces based on the Cassie-Baxter

regimen. Besides possessing basic self-cleaning properties, advanced self-cleaning surfaces

also contain optical functions such as transparency and images without aberration. Therefore,

an advanced optimization modeling will be established to recognize the equilibrium between

the self-cleaning performance and the optical performance based on the developed

mathematical models of two designed structured surfaces in ultra-precision machining. The

results generated from the optimization models will be compared with the experimental

results in order to validate models developed in the future.

The present experimental work has revealed that machined surfaces on a hydrophilic and a

hydrophobic material with micro patterns can achieve composite solid-liquid-air interfaces

when the scales of the machined pattern fall into a critical range. Experimental results indicate

that a hydrophilic material (i.e. PMMA) and a hydrophobic material (i.e. COC) can achieve a

high static contact angle if a micro pattern with appropriate surface geometries is machined on

its surface. The optimum width to depth ratio of the surface geometries in a frustum ridge has

been identified and proven through studying the wetting transition. It is expected that the

successful development of the specified models will significantly enhance understanding of

self-cleaning and accelerate the development of self-cleaning surfaces.

Keywords: Ultra-precision machining

6

Project Title: Microwave Processing of Titanium-Based Implants for Bone

Tissue Engineering

Name of Student: Choy Man Tik Dickson

Degree: PhD


Co-supervisor(s): Prof. William W.J. Lu (HKU)


Office: EF403


Metallic materials play an important role in the application of biomaterials to assist with the

repair or replacement of bone tissue that has become diseased or damaged. Annually, more

than a million patients worldwide receive treatment for the replacement of damaged hip and

knee implants. The number of replacements is continuous and will rise in such a way that a

very high demand for implant manufacturing is expected in the future. The materials used for

orthopedic implants, especially for those used in load bearing applications, should possess

excellent biocompatibility, superior corrosion resistance, high strength and low modulus, high

fatigue, high ductility and be without cytotoxicity.

Titanium (Ti), a promising metallic biomaterial, is suitable for load-bearing applications due

to a combination of high strength to weight ratio, high mechanical strength and good fracture

toughness. However, the existing fabrication methods involve high temperature, long

processing cycle time and high cost. The purpose of this research study is to develop a

microwave sintering method to fabricate Ti-based composites with appropriate mechanical

properties and shape memory effects for bone tissue engineering. The proposed fabrication

method would significantly reduce the sintering time and the processing cost while the

sintered part could fulfill the requirements for bone implantation.

The preliminary results have demonstrated that the microwave sintering method can

significantly reduced the sintering time from 200 min to 2 min. The reduced use of a

protective gas resulted in the reduction of the processing cycle time and costs while the

sintered specimens had good shape retention and no contamination after sintering. Both solid

and porous titanium samples have been successfully fabricated using different sizes of initial

powders and a temporary space holder. Macropores and micropores with rough wall surfaces

were found to be uniformly distributed within the sintered specimen which enabled the bone

tissues to penetrate the specimen for bone growth and to allow for body fluid transportation.

In order to understand the exact mechanism of microwave energy absorption by metallic

particles and the nonlinear relationship of the interaction between the microwave and the

heated materials, an analytical model will be derived. Based on the coupling relationship

between the electromagnetic field equation (Maxwell's equation) and heat transportation, the

analytic model will attempt to represent and describe the sintering conditions of a multimode

cavity microwave furnace, to assist in controlling the sintering process and to predict the

sintered behavior.

Keywords: Powder metallurgy, Microwave sintering, Titanium implants, Bone tissues engineering

7

Project Title: Wear Analysis of Diamond-coated Tools in Precision Machining

of Difficult-to-cut Materials

Name of Student: Fung Kai Yin

Degree: MPhil




Office: DE406


Precision machining of difficult-to-cut materials (e.g. silicon and SiC), which are used for

optical and micro-electro-mechanical (MEM) systems in severe environments, has received

increasing research interest in recent years. Precision machining processes, such as

single-point diamond turning (SPDT), are used to fabricate complex, freeform and mirror

finished surfaces and micro-patterns in a single pass. A brittle-ductile transition phenomenon

takes place, and ductile regime cutting at the nanometre scale of hard and brittle materials is

amenable to precision machining using diamond tools. However, quick tool wear and a short

tool life have been reported due to high hardness and sudden change of cutting mode and

cutting forces across different lattice orientations during precision machining of

difficult-to-cut materials.

Diamond coating is an emerging technology, by which the surface of a metallic surface can

be enhanced by increasing its corrosion resistance, thermal conductivity, and hardness.

Binderless diamond coatings, such as generated by chemical vapour deposition (CVD), on

tools have properties very close to or better than those found in conventional polycrystalline

diamonds. Diamond-coated tools have been considered as a potential candidate for precision

machining of certain metallic alloys. Though significant works have been done to investigate

the wear of diamond-coated tools in machining conventional materials, there has been no

open literature has reported wear analysis on diamond-coated tools in precision machining of

difficult-to-cut materials.

This project aims to characterize the wear mechanisms of diamond-coated tools in precision

machining of difficult-to-cut materials. Compared with the wear of single-point diamond

tools, the wear of diamond-coated tools is further enhanced by the adhesive strength between

the diamond film and tool substrate. Different thermal expansion and residual stresses

between the two materials can weaken the bonding of the diamond coating and cause

delamination. Several scholars reported high localized temperatures during nanometric

cutting due to nanofriction as well as shear heating at the shear zone. In addition, the

localized temperature enhances the carbon diffusion graphitization of diamond that results in

a shorter tool life. The delamination is also caused by abrasive and adhesive wear as well as

the growth of micro-grooves. Since precision machining requires a nanometric edge

sharpness of the tooling, the wear analysis of the diamond-coated tools is significant for

evaluating its applicability in the precision machining of difficult-to-cut materials.

Keywords: Tool Wear, Diamond Coating, Precision Machining


8

Project Title: Magneto-caloric Effect of Fe-based Metallic Glasses at Room

Temperature

Name of Student: Guo Daoqin

Degree: MPhil

Chief Supervisor: Prof. K.C. Chan



Office: DE406


Magnetic refrigeration, based on the magnetocaloric effect (MCE), has already been used in

low temperature applications. Over the last decade, many researchers have further explored

replacing conventional gas expansion/compression cooling techniques for room temperature

applications by magnetic refrigeration. As working refrigerants, the magnetocaloric effect of

magnetic materials is critical to the performance of a refrigerator. It has been reported that the

refrigeration capacity (RC) of amorphous materials is generally higher than that of crystals.

Among various amorphous MCE materials, Gd-based materials have attracted much research

interest due to their large magnetocaloric effect. However, their Curie temperature is still too

low for room temperature magnetic refrigeration. On the other hand, although Fe-based

amorphous materials have the advantages of high Curie temperature and low cost, their MCE

is much lower than that of Gd-based materials. In the present study, with the aim of

enhancing their MCE, the influence of Co addition to a Fe-based amorphous material was

investigated. Fe76-xCoxSi5Cr4Zr5B10 (x=0, 2, 4, 6) ribbons were fabricated by the

melt-spinning method, and their MCE was studied under a maximum field of 1.5T using a

vibrating sample magnetometer. It was found that the Fe74Co2Si5Cr4Zr5B10 ribbon performs

the best when both the refrigeration capacity and the Curie temperature are taken into

consideration (Tc=295K, RC=50J/kg). Although the RC value of the alloy is still not very

high, its Curie temperature is close to room temperature.

In order to improve the MCE, small amount of Gd (1%, 2% and 3% respectively) were

added, based on the previous results. The effect of the addition of Gd was characterized.

Although the Curie temperature dropped dramatically, the refrigeration capacity was

improved as expected.

In future work, it is envisaged that with further improvement of its RC, this material is

promising for magnetic refrigeration at room temperature.

Keywords: magnetic refrigeration, magneto-caloric effect, metallic glasses, Fe-based materials

9

Project Title: Chemophysical Analysis of Composite Materials Due to Erosion

Damage

Name of Student: Huang Wenfei

Degree: MPhil

Chief Supervisor: Dr. Gary Tsui

Co-supervisor(s): Prof. C.Y. Tang


Office: EF403


The polymeric Microsphere, as a parenteral drug delivery system is primarily developed for

the sustained release of drugs for prolonged systemic therapeutic effects, with subcutaneous or

intramuscular administration. The polymeric microsphere is mainly made from biodegradable

polymers and is usually in a free flow powder form with a diameter ranging from ten microns

to several hundred nanometers. A release of the drug from the microsphere is essentially

controlled by diffusion and polymer erosion, which largely depend on both the structure of the

microsphere and the polymeric matrix chosen for encapsulation.

There are several techniques the have been developed for the preparation of the polymeric

microsphere, including spray drying, single/double emulsion, polymerization and phase

separation. Spray drying was selected to be the technique for microsphere preparation in the

present research. It is because this method not only offers a more economic technology for

scaling up vs. other methods, but also can process both heat-resistant and heat sensitive

materials and provide high precision control over the microsphere structure, such as the

particle size, bulk density, degree of crystallinity and porosity.

In the present research, it would be amongst the first attempts in applying ultrasound to

atomize a polymer solution rather than using conventional high pressure air. The ultrasonic

spray method can be used to produce micro droplets with a diameter of less than 10 microns

and generate nanospheres that are out of the capability of conventional methods. In addition to

using conventional heating resource-hot air in the spray drying, other alternatives such as

microwaves, UV light and novel heating sources would be studied in this research. This is

because microwave can provide a homogenous heating effect in a shorter time than hot air,

which may help produce a microsphere structure with more uniform particle shapes. For the

UV light, it may provide a new approach to make a multilayer or well-encapsulated

microsphere with a cross-linked polymer shell.

The results of the proposed research work above will provide an innovative approach to

fabricate polymeric microspheres with a specific structure that cannot be produced by

conventional approaches, and would be benefit in controlling the drug release rate from the

microsphere. The project results will also contribute to the development of versatile structures

of the microspheres tailored for different applications and as fillers for composite materials.

Keywords: microsphere, spray drying, ultrasound, biodegradable, polymer, drug delivery

10

Project Title: The Research of Ultra-Precision Light Field Measuring

Microscope

Name of Student: Li Da, Adam

Degree: MPhil


Co-supervisor(s): Prof. W.B. Lee; Dr Sandy To


Office: DE404


The geometric complexity and high accuracy requirement brings many difficulties to the

measurement and characterization of machined Ultra-precision micro-structured surfaces.

Although many ultra-precision coordinate measurement instruments have been developed in

recent years for the measurement of complex surfaces, these measurement instruments fall

into the traditional off-line measurement process which requires the manufactured workpieces

to be removed from the machine tools to a separate measurement area. This is a very

challenging task for workpieces with relatively large dimensions and heavy weight. Currently,

the measurement is mostly based on assessment of the optical quality of the replica surface,

which is however an indirect measurement process which lacks efficiency and measurement

traceability. Therefore, it is desirable to directly measure the drums during the manufacturing

process. This can be achieved by on-machine measurement techniques.

This project aims to develop an on-machine Metrology System for performing non-contact

measurement of optical freeform surfaces with sub-micrometer form accuracy and surface

finish in the nanometer range, based on light field theory, and optical and image processing

technology.

The light field is a function that describes the amount of light faring in every direction through

every point in space. By inserting a microlens array into the optical system, one can capture

the light fields of a workpiece in a single photograph, which has the ability to create focal

stacks from a single photograph of a workpieces.

According to Fourier Optics, PSF is a system's impulse response of a focused optical system.

The image of a complex object can then be seen as a convolution of the true object and the

PSF. Accordingly, the object information can be achieved by the deconvolution of the image

and PSF. In this sense, applying 3D deconvolution to these focal stacks, we can produce a set

of cross sections, which can reconstruct the 3D information of the Ultra-precision

micro-structured surfaces.

The methodology used in this project can provide non-contact measurement in the

manufacturing process, combining a non-contact measuring apparatus with the freeform

surface characterization method into an integrated on-machine metrology system. It can

precisely reconstruct the surface appearance digitally. In this sense, it would be convenient to

proceed by following measurement steps, such as comparison, error analysis or evaluation.

Keywords: Light field, Micro-lens array, 3D information reconstruction, deconvolution,

ultra-precision measurement

11

Project Title: An Investigation of Processing Technology for Ultra-precision

Roller Embossing of Optical Micro-structured Surfaces

Name of Student: Lui Sing Yuen Jasper

Degree: MPhil




Office: DE404


Microstructure pattern are widely used in optical components, such as the light guide plate

for LED/LCD displays. The V-grooves on the light guide plate have function to guide the

optical path and scatter the light emitted from the light source based on the optical design.

The project aims to investigate the new method for ultra-precision machining of V-grooves

with patterned diamond tools. The objectives of this project are to:

i) Develop patterned single crystal tools with micro-structured rake faces for

ultra-precision machining;

ii) Study the mechanism of micro-cutting with the developed diamond tool and obtain

the optimal cutting parameters to improve the cutting performance;

iii) Produce microstructures on roller surface with the developed diamond tool and

investigate the rolling process for fabrication of the microstructure patterns on

plastic films by hot roller embossing;

iv) Study the effect of material bounce-back on the form accuracy of the hot-embossed

plastic film and the compensation methods in precision hot roller embossing.

The project applies patterned single crystal diamond tools with the shaped cutting edge of

several V-grooves, which will replicate the V-grooves on the workpiece surface within a

single cutting pass. The dominant factors affecting the form accuracy of microstructures

should also be studied as well: the tool wear, chip formation, material elastic recovery, etc.

Experimented analyses of the results include (i) the deviation between tool design and the

machined V-grooves structure, (ii) the swelling of the workpiece materials, (iii) the form

accuracy of the produced pattern, (iv) the cutting performance with variation of the

V-grooves angles and (v) the optimal settings of the machining parameters, etc. The Finite

Element Method can help in analyzing the compression and shearing, the distribution of

stress, strain and flow line tracing, material rebounce, force distribution, material

deformation for the rolling process.

The new fabrication method for V-groove back light plates, proposed by this project, not

only enhances the cutting efficiency but also provides a new method of production for the

back light guide plate with hot roller embossing technology, which promotes cutting

efficiency and reduces the production cost and time.

Keywords: Hot embossed rolling, micro-grooves

12

Project Title: Fabrication of Bioactive Titanium Oxide Coating on

Nickel-Titanium Using Plasma Electrolytic Oxidation

Name of Student: Siu Hin Ting

Degree: MPhil




Office: DE 404


When nickel titanium (NiTi) was introduced as a biomaterial because of its unique shape

memory and super-elastic properties, researchers began to focus their attention on the safe

use and apatite-forming ability of this material. Due to its high nickel content, nickel

inevitably is found on the surface of untreated NiTi. Therefore, surface treatment is necessary

to reduce the nickel content on the surface of NiTi to ensure the safety of the implantation.

Apart from the risk of nickel ions release that may cause allergic effects, NiTi is a bio-inert

and a poor osteoinductive material. The growth of body tissue on NiTi implants is not easy

unless a bioactive coating exists. In order to enhance the stability of implantation, a lot of

research work has been focused on the surface modification of NiTi.

This project investigated the feasibility of forming a thick and porous coating on NiTi by

plasma electrolytic oxidation (PEO) which is an effective technique that can be conducted at

room temperature. At such low temperatures, the bulk properties of the thermally sensitive

NiTi would not be affected. The Taguchi experimental design approach was implemented to

optimize the process parameters for PEO. Two L16 and one L27 Taguchi experiments were

conducted and optimized parameters in terms of the concentration of electrolytes, voltage and

processing time were found. The results from the Taguchi experiments indicated that an

alkaline environment was more suitable for conducting PEO treatment on NiTi than an acidic

one. Titanium oxide coatings of around 10μm thick with a porous structure were successfully

fabricated in a Na2SO4/NaOH electrolyte by an AC power source. XPS analyses showed that

the surface of the treated NiTi contained only a small amount of Ni when compared to the

substrate.

In addition, the influences of different parameters on the surface morphology, phase

composition and corrosion resistance were investigated. Crystalline titanium oxide coatings

with enhanced corrosion resistance were obtained. The immersion test results in a simulated

body fluid demonstrated significant improvement in terms of the apatite-forming ability of

the PEO-treated samples. Hydroxyapatite (HA) particles were observed on the treated NiTi

after 28 days of immersion whereas no particle was found on the bare NiTi samples. This

indicated that the PEO treatment developed in this study can improve the bioactivity of NiTi.

It is concluded that such a bioactive coating can reduce the nickel content at the implant

surface and enhance the apatite-forming ability, thus improving the performance of

orthopedic and dental implants.

Keywords: Niti, Taguchi method, corrosion, Plasma Electrolytic Oxidation

13

Project Title: A Multiscale Modelling Approach for Predicting the Critical

Undeformed Chip Thickness for Ductile Regime Cutting of

Brittle Materials in Ultra-Precision Machining

Name of Student: Xiao Gaobo

Degree: PhD




Office: DE406


Diamond cutting of brittle materials is an emerging technology that has great advantages over

traditional finishing processes such as grinding and polishing. It has higher machining

efficiency and higher form accuracy, and can be used to fabricate very complex surfaces such

as freeform surfaces. One of the key issues in the diamond cutting of brittle materials is to

determine the critical undeformed chip thickness for ductile regime machining. At present,

the critical undeformed chip thickness is mainly determined by costly and time-consuming

experimental techniques. Therefore, much effort has been devoted to predictive modelling of

the critical undeformed chip thickness. However, there remain some issues that have not been

well addressed by these models, such as the effects of crystal anisotropy and tool geometry.

In this study, a multiscale modelling approach, which is able to predict the critical

undeformed chip thickness in different crystal orientations, was proposed and verified by

taper cutting experiments using a typical brittle material, silicon carbide (SiC). The idea for

the proposed approach is based on the following assumptions. In the ductile regime cutting of

brittle materials, tensile stress exists behind the cutting zone. The tensile stress increases as

the undeformed chip thickness increases. Thus, the tensile stress will exceed the ultimate

tensile strength of the material when the undeformed chip thickness reaches a critical value,

which will lead to brittle fracture of the workpiece materials, resulting in brittle mode

material removal. Based on this assumption, the following approach was proposed to predict

the critical undeformed chip thickness for ductile regime cutting of brittle materials. First,

molecular dynamics (MD) simulations will be performed to calculate the yield strength and

ultimate tensile strength of the material in various crystal orientations. Second, the finite

element method will be adopted to model the stress distribution for different depths of cut.

Thus, the depth at which the tensile stress behind the cutting zone exceeds the ultimate tensile

strength can be determined. This depth is assumed to be the critical undeformed chip

thickness for ductile regime cutting.

According to the proposed approach, MD simulations of uniaxial tension and compression

tests were performed to calculate the ultimate tensile strength and yield strength of 6H SiC.

FEM was employed to model the stress distribution in the cutting zone under various depths

of cut. The results of the MD and FEM simulations were combined to predict the critical

undeformed chip thickness of 6H SiC for various crystal orientations. The modeling results

were then verified by taper cutting experiments. This is the first time that the effects of crystal

anisotropy on the critical undeformed chip thickness have been modelled effectively.

Keywords: brittle material, ductile mode cutting, critical undeformed chip thickness, multi-scale

modelling, ultra-precision machining

14

Project Title: Re-manufacturing of Aeronautical Components by Additive

Technology

Name of Student: Xin Ying Anita

Degree: PhD


Co-supervisor(s): Dr Stephen O'Brien (IC)


Office: W501e


What is adaptive machining? Basically, it is a technology that can adapt the machining

strategy to meet with complex manufacturing requirements. Adaptive machining not only can

offer a best-fit strategy within different constraints and align parts to a datum position, but it

can treat individual problems in each part, and make unique strategies for machining.

Adaptive machining can be widely used in aero-engine component repair and maintenance,

e.g. turbine blades. It is because worn blades have different geometrical structures and a

variety of defects, e.g. distortion, wear, and cracking. One single machining strategy cannot

solve problems with the same model. Adaptive machining technology may use a scanning

system, special CAD/CAM software and a computer connected to a CNC controller via LAN,

for real time process control. In this way, adaptive machining can automatically adjust the

programmed process to cope with subtle variations.

A great number of research and experiments have been made to improve adaptive machining

technology. These studies can be divided into four aspects, part geometry reconstruction, tool

path optimization, cutting force control, and rotary table and tool orientation optimization.

Parts surface recreation plays a significant role in adaptive machining. The process should be

efficient and accurate. Some researchers use 3D optoelectronic sensor devices and

touch-trigger probe inspection technique as the basis for the model reconstruction. For

example, a 3D non-contact optical measurement system is used to generate a welded blade’s

digitized polygonal model to restore its tip geometry. Research on tool path optimization has

been ongoing for decades. Many new tool paths have been developed in recent years, such as

the guide surface tool path, the iso-curvature tool path, the constant scallop tool path, etc.,

which have improved the quality of a part’s sculptured surface machining. Effective

machining force control can bring significant economic benefits, including improvements in

quality and reductions in cost. Most research managed to get process parameters on-line

(cutting force and the cutting tool’s coordinate positions) and adjust the machining force

according to certain algorithms. 5-axis CNC machines have two more degrees of freedom

than 3-axis machines CNC which offers many advantages in machining and brings new

research topics as well, e.g. dynamic performance of the machine tool and the rotary table.

All research topics mentioned above are not isolated and can benefit from the technical

development of adaptive machining.

Keywords: Adaptive machining, 5-axis, scanning, tool path, force control

15

Project Title: Modeling and Proactive Resilient Self-adaption of the Effect of

Tool Wear in Ultra-precision Raster Milling

Name of Student: Zhang Guo Qing, Peter

Degree: PhD




Office: DE406


Ultra-precision raster milling (UPRM) is an enabling machining technology which can be

used to machine non-rotational symmetric components with sub-micrometric or even

nanometric surface roughness. However, the occurrence of diamond tool wear during the

cutting process can definitely affect the surface finish of machined products. If diamond tool

wear makes the machined products unacceptable, the long previous cutting time and

parameter configuring time are wasted, lowering the productivity and effectiveness of

UPRM. Until now, a significant amount of research interest has been paid to the methods of

monitoring tool wear, both direct and indirect methods. Direct methods are carried out by

examining tool wear directly by optical microscope or scanning electronic microscope

(SEM). Indirect methods mean monitoring tool wear through indirect signals e.g. cutting

force signals, acoustic emission signals, vibration signals, power consumption etc.

In my research, an indirect tool wear monitoring method-monitoring tool wear by examining

cutting chips was adopted. This method is effective because during the UPRM process, the

diamond tool wear characteristics can be directly imprinted both on the machined surface and

cutting chips, and through checking the cutting chips, the tool wear can be examined. In this

method, the cutting chips are collected in an certain interval time and subsequently examined

by SEM, and the captured chip figures are used to establish the 3D tool wear model together

with cutting parameters. The machined surface quality can be predicted based on the 3D tool

wear model, and the machined surface quality can also be improved by changing the cutting

parameters so as to reduce the effect of tool wear on the machined surface quality.

This method can realize monitoring tool wear effectively without the need to stop the cutting

machine, and according to current experiments, it is found that the tool wear monitoring

methods proposed in this proposal are reliable.

Keywords: Tool wear monitoring, cutting chips, ultra-precision raster milling, surface quality

16

Knowledge & Technology Management

17

Project Title: A Study on the Relationship between Knowledge Management

and Organizational Performance in the Manufacturing Industry

Name of Student: Khan, Muhammad Aamir

Degree: PhD

Chief Supervisor: Prof. Eric Tsui

Co-supervisor(s): Prof. W.B. Lee


Office: EF403

Country of Origin: Pakistan

Knowledge Management is now a necessary organizational structural management system

with a large number of interrelated attributes. However, its three elements that are basically

found in the literature are: knowledge adaptation, knowledge distribution and knowledge

utilization. The knowledge management practices in firms depend on some basic practices.

One of the important ones for effective knowledge management is organizational culture.

Consideration of the practices of knowledge management is no longer limited to the

manufacturing industry; it is also very essential for different service sectors where the

performance, efficiency and effectiveness can be enhanced through the execution of suitable

knowledge management practices in line with the business strategy.

Several research studies examined how knowledge may be retained remain even after the key

persons have left the organization, and there have been a lot of studies in the implementation

of knowledge management and its assessment, but relatively little research has been carried

out on the relationship of knowledge management practices and organizational performance in

the manufacturing industries. Without continuously developing and diffusing organizational

knowledge, an organization would not be sustainable in the long-term. If an organization does

not succeed integrating the necessary skills, the performance will decline.

In this study, a better understanding of the different issues that exist in knowledge

management in relation to organizational performance will be developed. Performance

measurement scales will also be taken into account which relate to KM and identifying key

areas which help in determining organizational performance. Numerous detailed hypotheses

will be formed on the basis of Hong Kong MAKE Awards criteria which uses a Balanced

Score Card to show organizational performance measurements. The hypotheses later will be

used to generate research questionnaire to get definite data from manufacturing industry

which will help to analyze or develop different models for measuring KM performance in the

manufacturing industry.

The resultant hypothesis or models will be adequate to characterize the performance of the

different manufacturing industries to conduct or bench mark different KM practices among

different departments of an organization and later enhance their performance by analyzing the

feedback with respect to their planned targets.

Keywords: Knowledge Management, Organizational Performance, MAKE Awards, Balanced Score

Card

18

Project Title: Computational Organizational Narrative Generation (CONG) for

Decision Support Learning

Name of Student: Yeung Chui Ling, Charlie

Degree: PhD


Co-supervisor(s): Prof. W.B. Lee, Prof. Eric Tsui


Office: DE406


Decision making has long been a critical process in human life. Many researchers advocate that

the real-world narratives shared by experts or knowledge workers are helpful in teaching and

educating novices to learn new knowledge and skills. However, these narratives are valuable, and

limited, as the narratives are generated by the occurrence of incidents or expert domain

knowledge. In addition, the generation narrative by experts is time consuming and costly. It

results in limited narratives that can be used for decision learning. Due to the commencement of

the retirement tsunami in 2012, those highly skillful and well experienced employees with

valuable narratives in their minds have started to leave their workplaces. It means that

organizational knowledge is now regularly lost and less and less knowledge can be used for

supporting decision making in their organizations. In order to retain and manipulate valuable

narrative knowledge, it is important to develop a methodology to represent and analyze

narratives.

A literature review regarding narratives and narrative analysis has been conducted. It is found that

different types of narratives have their specific scaffolds and functions. However, it can be

summarized that a basic narrative is essentially composed of three main stages which are the

beginning, the middle and the end, as stated by Aristotle. Researchers also proposed narrative

analysis by adopting Labov’s model to investigate narratives in six dimensions. The conventional

approaches of those studies related to narratives were conducted manually by experts. The

production time and cost is high when there are a large number of narratives. Narrative

knowledge may not be retained and extracted in a timely manner.

This study attempts to design and develop a computational approach to conduct narrative

representation and analysis. The study focuses on investigating the planned and organized

narratives in written or oral versions. A narrative segmentation algorithm has been built in the

first part of study. It is capable of identifying and representing the narratives in terms of three

critical stages. In order to facilitate narrative analysis, a sentence restructuring algorithm has been

constructed to divide complex and complicated sentences in the narratives into simple clauses.

The narrative analysis algorithm will then be further developed to analyze the narratives by using

Labov’s model. The construction industry has been chosen as a reference site for trial

implementation of the new approach, and encouraging results were obtained after conducting a

case study.

The situation regarding knowledge loss in organizations has become severe since 2012.

Traditional approaches for narrative analysis and generation, which are time consuming and labor

intensive are inadequate for knowledge retention and decision learning. Therefore, the

computational approach to conduct narrative representation and analysis provides an important

means to facilitate humans to understand and assimilate the narrative knowledge.

Keywords: Computational Organizational Narrative Generation


19

Project Title: Assessment of Uncertainty in Quality of Knowledge in Research

Publications.

Name of Student: Farzad Sabetzadeh

Degree: PhD




Office: DE406

Country of Origin: Iran

Every day, there are many research shares and proposals submitted to journals or scientific

judgment panels for selection. The subsequent selection process carriers an important role in

the reputation or financial resources that are bound to these decisions. On the other hand, there

are also limitations in the number of research studies that can be accepted due to financial or

reputational reasons. This selection relies on the quality of knowledge that appears in research

proposals or is reported in research papers. While, in scientific disciplines, the reported

research works are expected to carry some specifications for recognition as valuable pieces of

knowledge, the assessment of quality and their subsequent selection are always prone to the

subjectivity of measurement that exists in the peer-review process. This can either originate

from the reviewer’s beliefs and assumptions about the quality of the proposed research (e.g.

Novelty) or the interpretation of the research work as the basis for the justification for a

decision about its selection (e.g. Validity).

My research aims to analyze the uncertainty level that exists in the propositional

knowledge quality assessment process in peer-review of scientific research works and aims to

propose an assessment framework to reduce the uncertainty in decision-making that is the

consequence of existing subjectivity.

This research will use the epistemological approach to the assessment of knowledge

quality. Using epistemological techniques, this study will develop a survey to measure the

belief level of the reviewers in a specific context (scientific field). Based on the belief levels,

justification for the decision is also drawn and compared accordingly (e.g. correlation,

divergence etc.). From this belief-justification pattern, an uncertainty model will be

developed to identify different locations for Justified True Belief (JTB) theory. This can also

be extended to different scientific contexts for comparison. This study aims to propose a

model that can help research assessment institutions achieve a higher confidence level in their

decision making process regarding various type of research works.

Keywords: Knowledge Quality, Uncertainty, Epistemology, Decision Making.

https://sixprd0111.outlook.com/owa/redir.aspx?C=8jSKjCd5NEKJGyoSEOEdyHOPichy0s8Ide3EWHZGuhNFsyxI9ECIl5n1FSdPidOsnHGF2_-S7h8.&URL=mailto%3afarzad.sabetzadeh%40connect.polyu.hk

20

Project Title: A Design Based Research to Conduct Knowledge Audit for

Unstructured Business Processes

Name of Student: Yip Yuen Tung

Degree: PhD




Office: CF404


This synopsis briefly presents a knowledge audit case study in Hong Kong, featuring the

application of a pattern-detecting knowledge representation methodology. Major outputs of

existing knowledge audits include knowledge maps and inventories which record the

knowledge flow and stock. While the nature of knowledge work and processes are becoming

complex, a knowledge representation methodology in a knowledge audit is to be evolved to

support learning and pattern-detection.

Design-Based Research (DBR) methodology was adopted in this knowledge audit research.

DBR is a systematic study of materials and sources in order to establish facts and reach new

conclusions through a series of iterations.

The developed knowledge representation methodology visualizes major activities and

knowledge assets involved in an unstructured process in a map. Observations from the map

were identified as team espoused theory, which was then compared with the design and

theory-in-use of the unstructured. The differentiation between the espoused theory and the

theory-in-use is regarded as an emergent pattern.

This research opens us a new gateway in the knowledge audit literature, exploring the

relationship between the knowledge audit and pattern detection. It aids organizations to

navigate in unstructured and complex business landscapes. The knowledge representation

methodology developed in this research visualizes and analyzes the complex interplay

amongst agents, activities and knowledge in unstructured processes. Patterns revealed from

the knowledge representation methodology facilitate team learning, nurtured team

mindfulness and anticipation in unstructured work processes.

The knowledge audit research also solves the practical concerns of organizations in

anticipating emergent patterns in unstructured processes. This knowledge audit research

generates emergent patterns and weak signals with the consensus of the team and thus can be

strong navigator for future knowledge management initiative planning.

Keywords: Knowledge Audit, Knowledge Elicitation, Knowledge Representation, Unstructured

Business Processes


21

Project Title: Theoretical and Experimental Investigation of Nano-surface

Generation in Ultra-precision Freeform Polishing: Process

Modelling and Optimization

Name of Student: Ho Lai Ting, Lesley

Degree: PhD



Dr W.M. Chiu


Office: DE406


Computer-controlled Ultra-precision Fluid Jet Polishing (UFJP) is an enabling technology for

machining surfaces with sub-micrometer form accuracy and surface roughness in the

nanometer range. Polishing technology is commonly used to remove undesirable machine

marks. In recent years, UFJP has not only been used for removing tool marks in order to

achieve super finished surfaces but has also been used for controlling the form accuracy of

such surfaces. However, it is noteworthy that UFJP can obtain a stable and controllable

Gaussian-profile of the removal function with no tool wear and sensitivity of the polishing

position. Although structured surfaces are commonly produced by laser, etching, plasma and

other technologies, UFJP shows a great potential with regard to its application value in the

generation of 3D-structured artifacts and patterns for 3D-structured surfaces of

difficult-to-machine materials.

Three-dimensional (3D) structured surfaces have been widely used in different applications

such as self-adhesive sensors, compound lenses in phonics products, etc. UFJP is an emerging

process which possesses the technological advantages of localized force and less heat

generation, as well as a stable and controllable material removal function without tool wear.

Generally, the workpiece material is removed or deformed by shear stress and pressure caused

by the fluid jet. Due to the complex machining mechanism, up to present, there is still a lack

of deterministic models that have been developed with consideration of all the operational

parameters so as to predict the material removal rate accurately. This study presents a novel

technology for the fabrication of 3D-structured surfaces made of hard and difficult-to-machine

materials such as glass by using UFJP. As a part of my PhD project, study of the modeling and

simulation of structured surface generation by using the FJP process is presented. The

polishing mechanisms of FJP are firstly explained, and then the material removal rate (MRR)

of FJP is derived. Hence, a structured surface generation model is proposed and explained. A

series of experiments have been undertaken, and the results are discussed.

Keywords: Ultra-precision Fluid Jet Polishing (UFJP), polishing material removal models,

3D-structured surfaces, surface generation


22

Project Title: Intelligent Reporting of Intellectual capital for Value Creation in

Knowledge-intensive Organizations

Name of Student: Cai Linlin

Degree: MPhil




Office: DE406


Since the world’s economy has rapidly changed from industrial to knowledge based, the

difference in the market-to-book ratios demonstrates that the current financial accounting

systems in practice are not adequate definers of economic value or resources. This difference

is considered to be” Intellectual Capital (IC)” from which value emanates. Therefore, a

rapidly increasing interest in and understanding of the role that intellectual capital resources

play in the work place are focused on. To manage IC, the IC must be located in the

organization first. The internal reports, related interviews, surveys, guidelines, as well as the

space of social media are all places in which IC exists. However, the dynamics and

complexity of IC became the big challenge for organizations to extract IC from these massive

amounts of materials.

The conventional method of extracting IC-related information is manual content analysis.

This labor-intensive method is greatly affected by personal bias in terms of the coding

process, which is used in practice. Automatic extraction that is assisted by computer offers

great help to cope with a huge volume data. However, the barrier in ignoring IC keywords’

context decreases the accuracy of IC, and should be recognized. Thus a comprehensive IC

keywords taxonomy is established to bridge this gap here.

(1) Identify actual IC reports from knowledge-intensive organizations that can be found.

(2) Do a manual content analysis for keywords commonly used to identify IC and map with

the most appropriate IC category (Human capital, Structural capital, Relational capital)

(3) Search for these words and phrases in annual reports and count the instances.

(4) Establish the IC taxonomy

Then the IC taxonomy will be used to extract IC from the documents that stakeholders use to

make decisions. Finally, an interview will be conducted to test the relevance of extracted IC in

terms of creating value for the organization. The results will also be compared with manual

and existing automatic methods.

Firstly, this intelligent method greatly increases the volume of the documents that can be

mined, which enables IC-related information to be identified in time. The second significance

of this method is that the context of the IC is known through tagging the keywords, which

cope with the coding difficulty in terms of classifying IC. Thirdly, the inter-relationship

among the different components of IC is extracted. Thus the Intellectual capital reporting can

be produced in time and accurately; the dynamic intellectual capital reporting will enable

organizations to find internal managerial problems and external opportunities in a more

practical level.

Keywords: Intellectual Capital


23

Project Title: A Study of Process Optimization and Nano-surface Generation in

Ultra-precision Machining of Precision Rollers for Advanced

Optics Manufacturing

Name of Student: Mak Chung Hong, Markson

Degree: MPhil


Co-supervisor(s): Dr Sandy To


Office: DE406


Plastic films with embossed micro-structured patterns and specific optical properties are

become extensively use in a wide range of applications, including LCD TVs, cell phones,

computer screens and industrial controllers. Because of the capability of producing well

distributed patterns in one pass, the current manufacturing method for these kinds of products

mostly uses injection moulding. However, injection moulding contains some problems

especially on producing large size LCDs. Based on this reason; a rolling method instead of

injection moulding is presented.

On the fabrication of the rolling machine, the critical part is the roller drum with

micro-patterns. In the present study, the manufacturing process of precision rolling by

ultra-precision machining is firstly investigated. Hence, studies of the factors affecting the

surface generation and the dimensional errors of the precision roller by ultra-precision

machining are being undertaken. A process optimization model for ultra-precision machining

of precision rollers for advanced optics manufacturing will be built.

Currently, a series of experiments have been designed and undertaken for preliminary studies

on a Nano form 350 machine. The results show that the length to diameter ratio, the material

chosen, the position of the machining area, the number of cuts, the cutting mechanism, the

cutting strategy and the tool wear seem to affect the generation of the V-grooves. Currently,

some further experimental work is being planned and will be undertaken to study the effect of

the optimization strategy for the surface generation of microstructure surfaces by

ultra-precision machining of the micro-patterns on the precision rollers.

Keywords: micro-structured pattern, ultra-precision machining, optimization


24

Project Title: A Computational Organizational Modeling and Simulation

(COMS) for Technology Assessment and Forecasting of

Technology Intensive and Innovative Enterprises

Name of Student: Cheng Mei Na, Meina

Degree: PhD




Office: DE404


Changing technology, driven forward by innovation, affects everybody's business. Smart

organizations do not wait for change to happen but proactively monitor and take advantage

of changing environments and new innovations. On the whole, the existing methods help

technology management professionals considerably. However, they have a number of

limitations which include:

(i) Effective use for managing the future of technology: Many technology forecasting

methods are one-off, ad-hoc and time-consuming which rely heavily on expert

experience and opinions. The difficulty for maintaining up-to-date technology trends is a

major barrier to their effective use.

(ii) Future-oriented nature: impacts of new technologies cannot be easily predicted until the

technology is anticipated, extensively developed and widely used.

(iii) Lack of technology development simulation: The existing methods provide a reactive

feedback based on the input data instead of providing a proactive manner which is able

to simulate different stages of the technology development.

In this research, a computational organizational modeling and simulation system (COMSS)

will be established based on information retrieval and extraction, data and text mining,

statistical analysis and narrative simulation. The COMSS consists of three main modules

including technology intelligence, scenario planning and technology road mapping. The

technology intelligence module will be used for capturing and delivering the technological

data and information as part of the process whereby an organization develops an awareness

of technological threats and opportunities. A technology intelligence module will be built as

a significant data bank which is used to store the technological data and information as well

as to retrieve specific technology intelligence for scenario planning module use. In the

scenario planning module, the technology intelligence is then processed (i.e. filed,

rearranged, integrated) to provide a series of informed and plausible scenarios about the

possible future technology trends. According to the results of the informed and plausible

scenarios, the technology road mapping module will be used to assess the possible future

technology trends, identify the impacts of the changes in technology and market needs, in

terms of potential threats and opportunities, especially for disruptive technologies and

markets.

The proposed system is to enhance the stakeholders’ decision making process in identifying

new business opportunities for exploiting new technology, and exploring, assessing and

planning innovation opportunities which aims for further growth and development of

technology-intensive and innovation enterprises. The development of the integrated platform

will enhance the automation of technology assessment process which will not only save time

and manpower resources but also enable an organization to keep pace with the knowledge

cycle in technology innovation and compare itself with other organizations in the industry.

This is particularly important when managing R&D activities and strategic planning for

technology management so that the enterprises can remain competitive in the global market.

25

Project Title:

Post-adoption Behavior for Personal Learning Environment &

Network

Name of Student: Tsui Lai Na Miriam

Degree: MPhil


Co-supervisor(s): Dr Eric Seeto


Office: DE406


The 21st century is a Knowledge Work era. Highly unstructured and fast-changing working

conditions and information overload are characteristics of this era. Problems that knowledge

workers have to tackle may be novel and they have to search for information to support

decision-making. On the other hand, so much information is available, making it difficult and

time-consuming to find the right information and to manage it. These conditions require

knowledge workers’ ability to learn effectively and efficiently in order to be competent.

Personal knowledge management and learning is needed as it can help knowledge workers

become more productive. During the current decade, changes in technology provide a variety

of tools for knowledge workers to develop their own learning environment and network

(Personal Learning Environment & Network, PLE&N in short). Individual knowledge

workers have the freedom to choose a set of tools as their learning environment and build-up

networks for co-learning, connecting with people and finding expertise and so on.

Knowledge workers adopt a variety of tools for varying periods of time. Some tools are

adopted and used continuously for a long period of time while some are adopted and

discontinued soon. It would be interesting to study what influences them in this post-adoption

behavior. However, research on post-adoption of technology is relatively sparse compared

with pre-adoption (initial adoption). It is argued that the initial stage of adoption of

conventional IT products and services, which are largely authorized by organizations, is of

higher importance as they usually incur a large amount of initial cost in acquisition but

minimal operating cost thereafter. A large stream of research is on organizational adoption. On

the other hand, many of the PLE&N tools are used at individual and personal levels, and

adopted in a bottom-up approach. Evaluation on the tools or studies on the post-adoption

behavior are generally not supported by organizations. While there is a need to select the right

tools to support work and continuous, active and lifelong learning, individual knowledge

workers are generally not forced to use a certain tool, so it is worthwhile to study the factors

that influence post-adoption behavior and the reasons why some tools are continued or

discontinued in use.

This research study aims to study the post-adoption behavior for PLE&N tools on an

individual level. Different stereotypes of PLE&N users will be identified. Research

hypotheses will be built. A field survey will be conducted for some PLE&N tools which have

been deployed in the PolyU to collect data and test the proposed hypotheses. It is expected

that framework(s) will be built to explain the continuance and discontinuance behavior of

different stereotyped users. By understanding more on the post-adoption behavior, technology

providers can design better adoption and retention strategies, while tools deployers (e.g.

lecturers in PolyU) can better predict user behavior.

Keywords: Personal Knowledge Management, Personal Learning Environment & Network,

Post-adoption


26

Project Title: Development of an Knowledge-based System for Managing

Competitiveness and Market Leadership of Project and

Process-Oriented Organizations

Name of Student: Rozhkov, Mikhail

Degree: PhD

Chief Supervisor: Prof. C.F. Cheung

Co-supervisor(s): Prof. Eric Tsui


Office: DE404

Country of Origin: Russia

The main features of a competitive company in a modern global economy are innovation,

adaptability and a high level of performance. These features are influenced by employee

knowledge and competence. However, scientific competitive analysis based on knowledge

management has received relatively little attention. As a result, this project aims to develop a

knowledge management system with adaptive decision making capability for managing

market leadership (competitiveness) of project and process management oriented

organizations. This system will provide the best organizational competence and performance

for any task context.

The research will examine “cause and effect” relationships between a combination of

workplace and team characteristics on the one hand, and individual employee competence, job

satisfaction and performance, on the other.

Workplace characteristics are divided into organizational culture and organizational climate.

These factors are important because they manifest the environment (context) for people

activities and communication during the project. Team characteristics are considered as the

characteristics of the manager and employees (cultural values, competency profile, etc.).

Examination of the interactions between these two sets of characteristics contributes to a

significant variety of dependent variables: employee competencies, job satisfaction and

performance. It is assumed that employee competencies, satisfaction and personal

performance are the major contributors to the overall performance of the innovation project

team. Employee competencies are considered as the underlying characteristics of a person

(skills, knowledge, behavior, habits) that have a causal impact on effective performance in a

job. These competencies provide a basis for differentiating the best and average performance

and predict job performance with high probability.

Major outcomes made during the last year were the theoretical background for studying

concepts and methodological aspects. During the next year, the pilot study and the main data

collecting will be conducted. During the data analysis stage, various statistical methods and

data mining techniques are proposed. The most promising are Analysis of Variance (ANOVA),

Multivariate Analysis of Variance (MANOVA), Multiple Regression Analysis, Clustering,

Association rules and Bayesian networks. Hence, a knowledge-based system will be built for

supporting project and process oriented organizations to manage team competency. The

performance of the system will be validated through a series of trial implementations in

selected reference sites.

Keywords: knowledge management, competency, performance


27

Project Title: Intellectual Capital and Value Creation – A Paradigm Shift?

Name of Student: Mariza Tsakalerou

Degree: PhD

Chief Supervisor: Prof. W.B. Lee



Office: DE406

Country of Origin: Greece

Intellectual capital (IC) –from intellectual property and patents through staff technical skills to

relationships and networking with customers– has been identified as a critical business success

factor. The general consensus is that effective management of intangible (intellectual or

knowledge) assets within an enterprise often serves as a source of competitive advantage and

hence of value creation for the organization. The causal relationship between IC and

organizational value has been the subject of significant academic research, however, with

mixed results. While the majority of the studies demonstrate that IC is positively and

significantly associated with organizational performance, a non-negligible minority claim that

the linkage has been generally weak, if at all present. In a few extreme cases, it was found that

a company’s IC, surprisingly, had a negative impact on its financial and market performance.

This mixed picture is due to the fact that there is no widely accepted definition of IC. In fact,

in almost all definitions of IC all irrelevant intangibles (i.e. those that are assumed to have no

association with the firm’s future potential) are excluded; this renders the weak linkage even

more problematic. On the other side of the equation, the concept of organizational

performance is equally vague, ranging from the narrow (use of financial indicators) to the

broad (inclusion of non-financial, operational indicators). The situation is exacerbated by the

fact that measuring IC variables is difficult and the objectivity of the information is often

doubtful.

The proposed research is based upon the premise that IC is a complex phenomenon of

interactions, transformations and complementarities. None of the intangibles of IC (resources,

capabilities and competences) is sufficient per se for truly gauging successful performance.

The proposed research seeks to re-examine the relationship between IC and value creation by

taking into consideration all relevant factors and by attempting to identify the interplay

between all the variables involved. In this sense, IC variables (intangible assets) will be

studied in tandem with financial (tangible) assets and knowledge management practices.

Furthermore, organizational performance will be defined broadly to include not only financial

indicators but also competitiveness and human factors.

Finally, the patterns of IC definition and usage across different business sectors (such as

services vs. manufacturing), company sizes (such as SME’s vs. MNE’s) and national origin

(such as European vs. Asian firms) will be identified and classified carefully. This will be

established through a meta-study on the behavior of the service and manufacturing industries

regarding the importance of intellectual capital on company performance. The results of the

comparative meta-study of published articles on the subject will be tabulated and presented for

each industry sector respectively, so as to provide a sound empirical basis to compare and

contrast them. Preliminary data indicate that doing so will broaden the perspective of this

research by observing multiple new dimensions of IC and will possibly lead to a paradigm

shift for the IC vs. company performance issue.

28

Project Title: Modeling and Ultra-precision Machining Micro-functional

structures for Heat Exchanger

Name of Student: Wang Haitao

Degree: PhD


Co-supervisor(s): Dr Sandy To


Office: DE404


The research is dedicated to the theoretical investigation of micro-functional structures for

realization of heat transfer enhancement. The research includes (i) the numerical simulation

of the micro-structure for heat transfer (ii) the ultra-precision machining of novel

micro-structure patterns (iii) the experimental work on CPU coolants with the micro-structure

heat exchanger.

First, the simulation models for heat transfer have been developed and various models have

been compared. According to dynamic characteristic and the thermal characteristics of fluid

flowing over the complex three dimensional micro-structures, the mathematical models have

been set up and the computational fluid dynamics (CFD) software have been applied to

simulate the effect of the micro-structure in the flow field. The parameters of the novel

micro-structure were optimized under the same boundary condition as the experimental

environment. The results show that the novel three dimensional micro-pyramid arrays have

good performance in heat transfer, without sacrificing an increased pressure drop.

Second, because the enhancement of heat transfer and pressure loss is very sensitive to the

roughness of the surface, ultra-precision raster milling was applied to manufacture the

surface. The process of diamond cutting has been analyzed and optimized.

Third, the CPU coolant experiment has been carried out to validate that the novel

micro-structures can transfer the laminar flow to turbulent flow, thereby enhancing the heat

transfer. Because this transfer will increase the friction drag, the balance between the heat

transfer rate and pressure drop becomes an important issue in designing the coolant flow

passages for the high flux heat removal encountered in micro-processor chip cooling. The

experimental work was compared with the results of numerical simulation the both results

matched very well.

Keywords: Micro-functional structures, Heat transfer enhancing, Numerical simulation,

Ultra-precision machining, CPU coolant experiment.

29

Project Title: Development of an Intellectual Capital Driven Knowledge Audit

Methodology with Application

Name of Student: Gu Jie, Jessica

Degree: MPhil


Co-supervisor(s): Prof. Benny Cheung, Prof. Eric Tsui


Office: CF405


The mining of important knowledge assets of an organization is both a time consuming and

often subjective process. A key task in this process is to critical assets components to the

business activities and goals of the company. This research tackles the shortcomings of

traditional knowledge asset audit methodology which operates on well-structured business

flow and process. These traditional methodologies are not suitable for knowledge intensive

companies which often have business processes that are highly unstructured, in which there is

no standard business flow. Knowledge assets of these companies do not merely reside in

structured forms, but are also embedded in many scattered and massive unstructured sources

such as emails, meeting records, publications, newsletters, websites, chat rooms, instant

messages, and blogs, etc. The aim of this research project is thus to develop a method to elicit

the knowledge asset items and assess their relative importance for subsequent knowledge

capture from the vast amount of unstructured information embedded in various sources of a

company. An intellectual capital (IC) framework which classifies the important intangible

assets of an organization into human capital (HC) , structural capital (SC) and relational

capital (RC) is adopted. A traditional IC value tree which shows the hierarchy and relative

importance of the chosen IC components (i.e., the subset of HC, SC and RC), is also used as

a framework to identify the knowledge inventory that is important or of relevance to the

company. However, conventional approaches in construction and elicitation of the IC

components are not only very time-consuming but also require the facilitation skills of

experienced IC management practitioners. Therefore, this research study offers an alternative

choice to companies to overcome such limitations. In this study, an IC oriented knowledge

elicitation system (iCOKES) is developed with a text mining algorithm to reveal the relevant

IC components. An IC thesaurus model is also built with an IC domain dictionary and

taxonomy to discover the important IC components that frequently appear in various

unstructured information sources. The IC components were extracted and relative importance

calculated according to their frequency of appearance in the various documents being mined.

An interview was conducted with the stakeholders to validate the IC value tree obtained from

the semi-automatic text mining. These IC components are presented in a structured IC value

tree which acts as a template for the searching and examination of knowledge activities,

knowledge inventory and knowledge flow in a company. Through case implementation in a

public utility company – The Hong Kong and China Gas Company Limited (Towngas), the

developed IC driven knowledge audit methodology has been executed and validated. The IC

driven knowledge audit methodology developed in this project has demonstrated to be a more

efficient approach to elicit IC items objectively with minimum amount of human

intervention or human bias in the construction of the IC value tree. The method has also

shown the capability to elicit IC components from a large amount of unstructured

information.

Keywords: Intellectual Capital, Knowledge Audit, Unstructured Information Management,

Knowledge Elicitation, Text Mining.

30

Logistics Engineering

31

Project Title: An Investigation on Closed-Loop Supply Chain using Priority

Based Genetic Algorithm Approach

Name of Student: Chen Yongtong, Cathy

Degree: PhD

Chief Supervisor: Prof. Felix T. S. Chan

Co-supervisor(s):


Contact Number: 5931-3915

Office: EF403


The closed-loop supply chain (CLSC) has become more popularity in recent years due to a

number of reasons. One of the most prominent reasons is that environmental issues have

gained increasing attention, and aiming another crucial reason for operating CLSC is the

cost.

The CLSC problem consists of transportation and facility location problems. First of all,

transportation problem (TP) was originally proposed by Hitchcock (1941) and became a

well-known basic network problem. The facility location problem received much attentions

since 1985 (Aikens 1985), aiming to decide the number of distribution centers and to find

good locations so as to satisfy customer demand at minimum facility operation costs and

delivery costs. Optimization of these problems is known to be NP-Hard.

In this academic area, most of these NP-hard CLSC problems are formulated into linear or

nonlinear problems. To solve this kind of NP-hard problem, an exact algorithm is convenient

but for large scale problems, the computational time is too long. Hence, the heuristic

algorithm, such as Genetic Algorithm (GA), becomes an efficient method and has gained

more popularity. Meanwhile, competition on the optimization quality in terms of solution

quality and computational time becomes one of the main focuses in the literature in this area.

In my study, a typical six-level closed-loop supply chain network has been studied. To deal

with the problem, a novel two-stage priority based Genetic Algorithm (GA) is developed.

This algorithm consists of two stages, decomposing the CLSP into two sub-problems. In the

first stage, the genetic algorithm is applied to generate a routing solution. After that, a cost

ranking heuristic is applied to determine the actual allocation quantity in stage two. The

result will give feedback into stage one to complete the solution. This novel algorithm

enhances the genetic searching ability of the GA in solving this kind of problem. To test and

demonstrate the optimization quality of the proposed algorithm, five numerical experiments

have been carried out. The results show that this proposed GA can get reliable and higher

quality results with shorter computational time compared with LINGO.

Keywords: Closed-loop Supply Chain, Genetic Algorithm, Reverse distribution, Linear Programming

32

Project Title: Integrated Planning of Berth Allocation and Quay Crane

Scheduling Problems

Name of Student: Ma Hoi Lam

Degree: PhD

Chief Supervisor: Prof. Felix Chan


Office: DE406


In maritime transport, vessels are carrying containers and travel from one container terminal

to another. When a vessel arrives at a container terminal, it will firstly wait at the harbor.

Once it receives a signal regarding its berthing position, it moors to the assigned berth, and

starts the loading and unloading operations. After finishing all the operations, the vessel will

leave and travel towards the next destination. To smoothly conduct these terminal operations,

different planning activities are needed, such as Berth Allocation Planning (BAP), crane

assignment and schedule planning, yard storage planning, etc. BAP is regarded as the first

step. It determines the berthing time and position of each incoming vessel. It is also well

recognized as the leading problem in terminal operations because it is not only directly

influencing the rest of the terminal operations but is also significantly influencing the

customer service level.

In term of business, competition among container terminals is getting more rigorous. In order

to survive in this environment, terminals strive to retain their customers by providing them

with good service quality. However, subjected to the facilities constraints, container

terminals may not be able to provide good service quality to every customer. In practice,

some customers are more crucial and critical to the terminal because they offer higher

container business volumes or have a long-term partnership with the terminal. In this

connection, a stable and higher service priority should be provided to this group of

customers.

In the literature, BAP is not a new research area. However, it is surprised that only a few

research projects studied the customer service level in vessel scheduling with consideration

of customer priority. In this study, a Genetic Algorithm (GA) is proposed to deal with the

above mentioned BAP problems. The proposed GA is capable of ensuring the service level of

the important customers with higher priority, meanwhile maximizing the service level of

other customers. In addition, the proposed GA also adopts multi-objective decision making,

so that more customers can be served by minimizing the transfer of vessels.

Keywords: Genetic algorithm, Service priority, Berth allocation, Vessel transfer, Terminal.


33

Project Title: Integrating Production Scheduling and Mold Maintenance

Planning: An Genetic Algorithm Approach

Name of Student: Wong Chun Sing, Sing

Degree: PhD

Chief Supervisor: Dr Felix Chan

Co-Supervisor(s): DrNick Chung


Contact Number: 3400 4778

Office: QR808

Country of Origin: China(HKSAR)

In general, production scheduling problems focus on determining the optimal work input

sequence to achieve the business goals. In today's competitive business environment, mold

manufacturing companies are facing a tradeoff between quality and the cost of the injection

mold. They tend to reduce mold costs, which may lead to lower quality and reliability of the

mold. Higher risk in production results, such as failure of the molds that causes idleness of

the injection molding machine. Maintaining high plant reliability in daily production is a key

issue for factory owners.

Injection molds are the major components in the operation of injection molding machines

and represent a significant share of the capital investment for plastics manufacturers. Thus,

they should be in good condition and always available for use. In many production

scheduling problems, important production resources such as injection molds are usually

assumed to be available without interrupting the production schedule. However, in reality, a

molds' condition is subject to deterioration relative to both of their usage and age, and thus

require maintenance or reconditioning which will disturb the normal production activities.

Although there are many studies related to maintenance and tool life management, there are

few papers that integrate mold maintenance into production scheduling. This combinational

problem is even more complex since mold allocation, mold age, and mold maintenance have

to be considered in the production schedule.

The objective of this study is to propose a simulation model with a genetic algorithm

approach to deal with this kind of combinational problem, by integrating mold maintenance

into production scheduling, aiming to minimize the makespan of the jobs. The significance

and benefits of integrating mold maintenance into production scheduling will be testified

through hypothetical numerical examples. The simulation results show that the proposed

integrated approach can improve the production performance with reduced makespan.

Keywords: Genetic algorithm, Production scheduling, Mold maintenance, Simulation

34

Project Title: Storage Allocation and Yard Trucks Scheduling in Container

Terminals using a Genetic Algorithm Approach

Name of Student: Wang Zhengxu

Degree: PhD


Co-supervisor(s): Dr Nick Chung


Office: DE406


With the development of techniques and trade globalization, marine transportation has played

a very important role in logistics networks. Container terminals are essential spots in the

networks, and the efficiency of container terminals is the most important factor which

influences marine transportation. The most vital factor of terminal efficiency in terminal

turnaround time which is the average time that a ship stays in a terminal. The study of

operations in a container terminal to shorten the turnaround time is quite necessary.

Container terminals put much effort on shortening the turnaround time by developing various

decision supporting technologies to optimize the terminal operations. In general, these

operations include berth allocation, quay crane scheduling, yard truck scheduling, yard crane

scheduling and storage allocation. When vessels come into a terminal, the operator should

first allocate berths and quay cranes for each vessel, and then allocate a number of locations

for discharging the containers. Finally, yard cranes and a fleet of trucks will be dispatched to

accomplish the corresponding loading and discharging operations. I focus on the storage

allocation and yard truck scheduling problem in this study. Storage allocation and yard truck

scheduling are two important problems that influence the efficiency of a container terminal.

Both problems have been individually studied by many researchers, but fewer people have

studied the integrated problem. Because of the intractability of the two problems, the

integrated problem is much more difficult to solve than the individual problems, and it has

troubled terminal managers for a long time.

This study proposes a new hybrid genetic algorithm with exhaustive heuristic and guidance

mutation to deal with the integrated problem of yard truck scheduling and storage allocation.

It is proven that the proposed genetic algorithm is effective in both small and large scale

cases by using a series of computational experiments. With the proposed hybrid approach, the

total delay time and total travel time of yard trucks are significantly reduced.

Keywords: Container terminal, Storage allocation, Yard truck scheduling, Genetic algorithm


35

Project Title:

Production and Scheduling in Supply Chain Management with

Uncertainty

Name of Student: Li Nan

Degree: PhD




Office: DE404


This project mainly looks at supply chain management with uncertainties and inaccuracies.

Simulation of the supply chain will be produced using mathematical methods. The

comprehensive production and distribution plan generated is expected to tackle the

uncertainty and fuzziness in supply chain management. To be specific, the project can be

divided into two major parts: supplier selection & order allocation and inventory

management.

Supplier selection and Order allocation: In recent years supplier selection and order allocation as an important part of supply chain

management, when facing unprecedented challenges and difficulties. High customization and

the fast changing market demands pressurize the modern supply chain management. The

problem is even more serious in Small to Medium Enterprises manufacturing (SMEs)

networks. The problem of how to form and coordinate manufacturing networks effectively

continues to form the basis of much research. A hybrid Fuzzy Analytical Hierarchy process

(FAHP) and Genetic Algorithms (GA) approach is presented in this thesis to address the

problem. This research is based on an industrial case study. Data and information on

suppliers are collected from a company acting as a system integrator in a SME manufacturing

network. The weights of the supplier in terms of both qualitative and quantitative criteria are

identified. Then, as a result of GA optimization, optimum combinations of suppliers and their

production tasks are determined corresponding to the requirement of orders and their own

capabilities. The results show that the proposed method is capable of optimizing the

configuration of manufacturing networks and provides visualized information for decision

makers.

Inventory management with uncertainty and inaccuracy:

Inventory in a key link in the whole supply chain management, while the successful control

of inventory would significantly reduce the cost of inventory and out-of-stock levels which

are important for the performance of the supply chain. However, the existing uncertainty and

inaccuracy in a supply chain, such as demand and inventory levels, makes the problem hard

to control. This part of the project, first of all, tries to simulate and model the supply chain.

An integrated control strategy is then produced and compared with traditional methods to

evaluate the effectiveness of the proposed methodology.

Keywords: Supplier selection, Order allocation, Inventory management, Statistic analysis, Demand

forecasting


36

Project Title: An Integrated Green Supply Chain Framework for Sustainable

Industrial Development

Name of Student: Zhang Shuzhu

Degree: PhD

Chief Supervisor: Dr Carman K.M. Lee

Co-supervisor(s):


Office: QR808


Environmental deterioration and resource consumption have promoted public concern over

sustainability and environmental issues. More and more researchers and practitioners are

considering the “green” influence in their research and business. The objective of this study

is to discover, integrate and evaluate the “green” factors in the supply chain area. In the

literature on the strategy, operation and performance of green supply chains have provided

significant insights about the importance and efficiency of green supply chain management

(GSCM) and business value for customers. For example, Sarkis (2003) introduced a strategy

decision framework for GSCM with an evaluation by an analytical network process (ANP).

Chen et al. (2012) proposed a new business strategy selection of GSCM from another

perspective containing different possible factors. Zhu et al. (2008) indicated a measurement

model for GSCM practices implementation. Most researchers design their frameworks from

their own empirical study. However, the validity of the mentioned dimensions, the

interdependency and interrelationship among these dimensions are still under explored. The

lack of executable-level practice from the environmental perspective is also a bottleneck for

the implementation of the GSCM system due to the fact that most of the previous studies

focus on the strategic level.

In this research, we will conduct a comprehensive study of GSCM and sustainable industrial

development. First, we will summarize and compare the proposed frameworks and

assessment tools from previous studies. Based on that, we will try to provide our GSCM

framework, which aims not only to help supply chain parties understand the importance of

environmental and sustainable issues, but will also assist them with a referable and

applicable system for their implementation. To illustrate the practical implementation of our

system, new green processes will be designed and developed with the support of the

information system, involving all the supply chain parties, such as suppliers, manufacturers,

distributors, retailers and even the customers. Within these business processes, various key

performance indicators (KPIs) can be improved significantly with the support of our

optimized solution.

Keywords: Green supply chain management, Sustainable development, Performance management,

Reverse logistics, Closed loop supply chain

37

Project Title: An Approach for the Perishable Product Logistics Based on

Real-time Monitoring with Radio Frequency Identification

(RFID)

Name of Student: Wang Lixing

Degree: PhD

Chief Supervisor: Dr Andrew Ip



Office: DE406


Perishable products are essential in our lives. Spoilage or contamination of these products

can lead to serious consequences. Transportation involving Third-Party Logistics (3PL)

companies is a weak point for product management, so it is necessary to improve the

management of perishable products.

This research aims to study the transportation management of perishable products based on

real-time monitoring of the entire supply chain using Radio Frequency Identification (RFID)

technology. The study proposes a system called the Monitoring-based Decision Support

System (MDSS) which integrates three modules for the major functions, namely a Real-time

Monitoring Module (RMM) with enabling RFID for product quality evaluation, a

Forecasting and Warning Module (FWM) for arrival time prediction and emergency warning,

and a Decision Support Module (DSM) for vehicle and emergency management.

In the MDSS, environmental factors, product information and the evaluation results from

RMM are transmitted to FWM for forecasting and warning judgments. If anything abnormal

occurs, the corresponding information is then transmitted to DSM for emergency

management.

RMM is designed with RFID technology and sensor networks. The module introduces a

hybrid algorithm that combines the k-Nearest Neighbor algorithm (k-NN) with Artificial

Neural Networks (ANN) to evaluate product quality. FWM applies Fuzzy Case-based

Reasoning (CBR) in its forecasting function, and a fast Rule-based Reasoning (RBR) in its

warning function. In DSM, an Improved Quantum-inspired Evolutionary Algorithm (IQEA)

and the Genetic Algorithm (GA) are applied to create an optimal schedule for vehicle

management before the transportation of perishable products. These two algorithms aim to

solve vehicle schedule problems in different scales. In addition to the static optimization of

the vehicle schedule, DSM can also help to cope with any emergency. Using heuristic

approaches, DSM adjusts the vehicle schedule and provides suggestions on how to cope with

any emergency.

Further, a particular case is studied to test the performance of the system. The results of the

case study show that MDSS has a positive effect on perishable product management

especially during transportation. For further research, the system can be extended to manage

perishable products in the entire supply chain, including storage, retailer and recall, in

addition to delivery.

Keywords: RFID, Optimization, Decision support system, Logistics, Perishable product

38

Project Title: Design and Optimization of RFID-enabled Wireless Sensor

Network (WSN) Monitoring System for Biological and

Pharmaceutical Products Supply Chain

Name of Student: Ng Chun Kit, Felix

Degree: MPhil

Chief Supervisor: Dr Andrew Ip



Office: DE404


Biological and pharmaceutical (B&P) products are essential and important at all times for

human beings. In current era, the demand and requirements of these products are

dramatically growing. The quality and integrity of these products has become the most

important concern. As a key factor for ensuring the quality and integrity, the ideal product

status monitoring system needs to be very sophisticated and comprehensive in order to help

the supply chain parties including manufacturer, supplier and distributor to take timely and

proper action to cope with mishandling and unpredictable incidents before an accident is

happened. However, most of the current monitoring systems of B&P products found in

supply chain industries cannot meet the requirement well. As a result, this research project

aims to develop a Radio Frequency Identification (RFID) enabled Wireless Sensor Network

(WSN) monitoring system for B&P product supply chain management which can provide

real-time and systematic product status information for better product management.

WSN, is an emerging technology in recent years which possesses a high capacity to be

applied in many applications of various areas such as military, environment, health, home,

and industry. However, adopting WSN technology is a challenging task. It requires

consideration for a series of design factors, including production cost, energy efficiency,

sensor nodes placement, sensing coverage, network connectivity and fault tolerance. These

design factors are interrelated and there are trade-offs between them. For example, many

applications require a high degree of sensing coverage, but this requirement needs a

relatively a large number of sensor nodes which also implies a higher setup cost. Thus, the

trade-off between the coverage requirement and the cost need to be considered. In order to

effectively manage the trade-offs between these factors, an optimization model is proposed.

The framework of the model contains three stages. In the first stage, a minimum number of

sensor nodes and relay nodes will be placed and fully cover the monitoring area. Next, the

sink of the WSN will be put in an optimal place so that the total transmission power of all the

relay nodes can be minimized. Lastly, the number of additional relay nodes is calculated in

order to improve the fault tolerance ability of the WSN. The calculation is based on several

constraints which include the power consumption density of the relay nodes, the mean time

to failure (MTTF) of the WSN and the connectivity between additional and original relay

nodes.

To further enhance the proposed model, RFID technology will be proposed to be integrated

in the model later to improve the product identification, trace and track abilities as well as

systematical product management.

Keywords: Radio Frequence Identification (RFID), Wireless Sensor Network (WSN), Supply Chain;

Optimization, Monitoring System

39

Project Title: A Decision Support System for Managing Performance of

Logistics Service Providers in Cross-Border Operations

Name of Student: Lam Hoi Yan, Cathy

Degree: PhD

Chief Supervisor: Dr K.L. Choy

Co-supervisor(s): Dr Nick Chung

Dr George Ho


Office: EF403


In recent years, many manufacturers have moved their production facilities from Hong Kong

to cities located beyond the Pearl River Delta (PRD) region. This allows them to enjoy lower

operating costs than if the operations were based in Hong Kong. The supply chain operation

mode for logistics service providers (LSPs) has changed from local to cross border due to the

geographical separation between Hong Kong and China. Under the present situation, it is a

challenge for the LSPs to provide good logistics service solutions to customers while

optimizing the use of their limited resources in the warehouse. This results in the need of

formulating an operations strategy with the decision support system. In order to enhance the

performance of a supply chain, the warehouse operations efficiency in order delivery should

be focused to shorten the time spent in handling the complicated decision-making process.

However, it takes time to make critical accurate decisions regarding resource allocation and

job arrangement, in particular in handling high value goods, when considering cross-border

requirements.

The purpose of this research is to design a decision support system to support the logistics

service providers in assisting the decision-making processes for managing warehouse

performance of cross-border logistics operations. The decision support system helps to

provide warehouse order picking and a packing plan by considering various customers'

demand and customs requirements, determine the appropriate resource allocation strategy

such as pallet and carton so as to increase the efficiency in handling an order, and determine

the order handling sequence to ensure the order fulfillment process can be achieved on time.

Therefore, a case-based reasoning (CBR) approach is proposed to formulate the order

picking plan by applying useful information from past order records to the new incoming

orders. A hybrid approach called the case-genetic algorithm-based decision support model

(C-GADS) is then formulated to assist warehouse delivery order planning in fulfilling

customer orders by taking into consideration customer requests, warehouse operations

arrangements and cross-border regulatory policies. The C-GADS model integrates a k-means

clustering algorithm and the genetic algorithm (GA) with the CBR technique in classifying

new customer orders into case groups with the highest similarity value, allowing for

effectively selecting the most similar cases among the group as the solution for the new

order. Through integrating the relevant order attributes, the CBR engine then suggests the

resources allocation and appropriate action plan. By implementing the proposed system, it

helps the warehouse manager to achieve (i) appropriate resources allocation, (ii) increase in

warehouse operation efficiency, and (iii) better customer satisfaction.

Keywords: Decision Support System, Warehouse Operations, Logistics Service Providers, Case-based

Reasoning, Genetic Algorithms

40

Project Title: A RFID-based Resource Allocation System for Garment

Manufacturing

Name of Student: Lee Kar Hang Carmen

Degree: MPhil

Chief Supervisor: Dr K. L. Choy

Co-supervisor(s): Dr Kris M.Y. Law

Dr George T.S. Ho


Office: EF403


Resources management has remarkable impacts on production operations in terms of both

productivity and efficiency. It is always an important issue for all manufacturing sectors.

Manufacturers are striking for better resource allocation not only to maximize profits, but also

to reduce production cycle times. If insufficient resources are assigned to a particular activity,

the entire cycle time will be lengthened. In contrast, if excessive resources are assigned, the

work in process inventory will be increased, resulting in higher inventory costs.

At the moment, the emergence of fast changes in fashion has given rise to the need to shorten

production cycle times in the garment industry. Faced with the challenges in the competitive

market, garment manufacturers are being urged to achieve effective and efficient production

resource allocation for their survival in the industry. Currently, there is a lack of standardized

approaches for effective production resources management in the garment manufacturing

industry. This may lead to inefficiency in production performance. Problems arising include

inaccurate resources planning and allocation.

Usually, decision makers determine the required level of resources by evaluating the technical

requirements of garments, subjectively. Since their decision making processes involve

concepts which are uncertain and vague, such as ‘‘long’’ and ‘‘short’’, an attempt is made in

this research to apply fuzzy logic for handling imprecise information for determining resource

allocation plans. In addition, Radio Frequency Identification (RFID) technology is adopted to

capture data which is useful for improving the intelligence associated with the fuzzy logic.

This research presents a RFID-based Resource Allocation System (RFID-RAS), integrating

RFID technology and fuzzy logic concept for achieving better resource allocation with

particular reference to garment manufacturing. To confirm the viability of the RFID-RAS, a

case study is conducted in a Hong Kong-based garment manufacturing company, which is

considered a good representation of the sector, to help manage its resource allocation process.

Results indicate that the proposed approach outperforms the conventional approaches with

better decision making in resources planning.

Keywords: Resource management, Resource allocation, Garment manufacturing, Fuzzy logic,

RFID

41

Project Title: A Data Mining and Optimization-based Real-time Mobile

Intelligent Routing System for City Logistics

Name of Student: LIN Canhong, Jason

Degree: MPhil

Chief Supervisor: Dr. K.L. Choy



Office: EF403


The striking development of E-Business in recent years has increased the importance and

burden of transportation and logistics in urban areas. With more demanding and time-sensitive

customer service requirements as well as the subsequent competition from other logistics

companies, the practitioners are facing a more challenging situation when conducting a

quick-response and on-time delivery service. Additionally, in the presence of congested urban

areas resulting from frequent traffic jams, deliveries are often delayed and thus significantly

degrade the level of customer satisfaction. The traditional transportation management system

fails to reschedule the dispatching plan in a dynamic traffic environment. The complex city

transportation network also seems to frustrate the drivers when they are seeking the best

transport plan. In the industrial practice of city logistics, there are two major vehicle

dispatching problems that logistics companies encounter in their daily operations. The first

problem concerns the pre-designed routing plan before performing delivery. This is often

based on the driver’s past experience and aims at the shortest path instead of the shortest

traveling time. The second problem is that even though an optimal plan in terms of minimum

traveling time is established, it cannot remain optimum as the traffic conditions vary as time

goes by. An alternative time-optimal route for vehicles to take instead of staying on the

planned road which has bad traffic conditions, is difficult to find because traffic information

about other roads is not known to the driver.

The problems discussed above call for a decision support system to provide intelligent

transportation solutions. Advances in new telecommunication and mobile technologies such

as global positioning systems (GPS), geographic information systems (GIS), traffic flow

sensors, and Smartphones, make it possible to use real-time traffic information to improve

service level, enhance the economy and energy efficiency of logistics. Reflecting the real-time

traffic conditions throughout the city, the system can therefore constantly update the

time-optimal routing plan during transportation so the goods can be delivered to the customers

as soon as possible.

To help the driver determine time-optimal routing solutions in order to avoid congestion

according to the real-time traffic flow, a Real-time Mobile Intelligent Routing System is

designed and deployed on drivers’ Smartphones to help in routing decision making. Data

mining techniques are employed to discover the routing patterns from past cases of routing

plans so as to generate case-based routing plans for the drivers. A metaheuristic is used to

undertake the optimization of a real-time optimal routing plan based on real-time traffic

information. A case study and computational experiments demonstrate the effectiveness of the

proposed methods in significantly reducing the traveling time.

Keywords: Intelligent Transportation System, Real-time Vehicle Routing, Data Mining,

Optimization, Variable Neighborhood Search

42

Project Title: Enterprise Supply Chain Planning under Uncertainties

Name of Student: Liu Hongguang

Degree: PhD

Chief Supervisor: Dr P. Ji



Office: EF403


About 60 percent of the NYSE average daily trading is executed by programed strategies.

Most of those strategies are based on statistical arbitrage. Roughly speaking, statistical

arbitrage is a long horizon trading strategy that generates riskless profit. There are a variety

of automated trading systems which commonly employ statistical methods, data mining and

artificial intelligence techniques. In recent years, high frequency trading is also considered as

a special aspect of statistical arbitrage as it utilizes many complex computing on time series

data. Some famous strategies like “Pair trading” and “Arbitrage merging” have been used to

generate profit.

However, few of those kinds of trading is done in the capital market of China, as there are

not enough financial tools to hedge the risk of the strategy. With the continuous development

of the financial market, especially the introduction of the Index future into the capital market,

more and more statistical arbitrage strategies have become available for mainland investors

and QFII. Yet whether statistical arbitrage opportunities exist in this specified market remain

unverified. Although many companies majored in programed trading were established in

mainland China during the last year, few generated significant profit.

In my research, firstly, a test for the profitability of a simple trend trading system will be

conducted using the tick data of the Index future of HS300. Multicharts—a highly automated

trading platform—will be used to optimize the related parameters. Secondly, I will

implement the test of the statistical arbitrage strategy proposed by S. Hogan, R. Jarrow and

M. Warachka on derivative market outcomes using the high frequency data of the Index

future. Finally, wavelet transformation will be introduced to form a more reliable short term

forecast, after that, a fresh new statistical arbitrage strategy will be built based on the results

of wavelet forecasting, and empirical tests will be conducted to verify the profitability.

The objective of my research is to clarify the market efficiency of the mysterious capital

market of mainland China, and at the same time, to develop an automated statistical arbitrage

trading system, whose parameters are dynamically optimized along with the changing of the

capital environments. Coded on the platform of multi-charts, this trading system could

possibly be used to generate profit.

Keywords: Forecasting, Wavelet, High frequency, Trading


44

Product & Process Design

45

Project Title: An Integrated Marketing and Engineering Approach to Product

Line Design

Name of Student: Ridvan Aydin

Degree: PhD

Chief Supervisor: Dr C.K. Kwong

Co-supervisor(s): Dr P. Ji


Office: DE 404

Country of Origin: Turkey

Product Line Design (PLD) is about the determination of the number of product variants and

their design specifications for satisfying the needs of different market segments.

Furthermore, PLD also aims to develop product variants for maximizing profit and/or market

share. Since PLD usually involves both marketing and engineering considerations, it is

necessary to integrate these two aspects for obtaining optimal PLD.

On the other hand, environmental issues, sustainability, energy efficiency, resource

consumption etc. are commonly considered in product design due to the environmental

consciousness of customers, government policies and international regulations. Therefore, a

number of qualitative and quantitative studies considering green supply chain management

(GrSCM) in product design have been published. However, the multiple dimensions of

GrSCM as green materials/components procurement, product-in-use, end-of-life

management and reverse logistics have not been considered in PLD in terms of energy

saving green technologies, reuse and recycling.

Discrete choice analysis (DCA) is a widely used choice modeling method used to capture

consumer purchasing behavior (preferences), considering tradeoffs among product attributes

for highly reliable demand prediction and market share estimation. Customer heterogeneity

and various uncertainties (data, model, time) are major concerns in adopting DCA in demand

modeling. In this project, fuzzy theory will be introduced into DCA to deal with the

uncertainty associated with survey data and thus a novel fuzzy demand prediction model will

be developed.

The objectives of my study are 1) integrating the multiple dimensions of GrSCM into PLD in

order to determine the product attribute settings for satisfying the needs of various customer

segments. 2) Estimating the market share of a new product considering the fuzziness of

survey data. 3) Developing a time-varying dynamic demand model for demand prediction

over a period of time with respect to new technology, and changes of market and customer

needs. 4) Minimizing the absolute error of the customer utility function for more accurate

market share estimation.

Keywords: Product Line Design (PLD), Green Supply Chain Management (GrSCM), Choice

Modeling, Market Share, Fuzzy Logic, Discrete Choice Analysis (DCA)

46

Project Title: Prognostics of Chromaticity State for Phosphor-converted

White Light Emitting Diodes Using an Unscented Kalman Filter

Approach

Name of Student: Fan Jiajie, Jay

Degree: PhD

Chief Supervisor: Dr Winco Yung

Co-supervisor(s): Prof. Michael. Pecht

(University of Maryland, College Park,

USA)


Office: DE404


Phosphor-converted white light emitting diodes (pc-white LEDs) have higher efficiency,

smaller size, lower power, consumption, and higher reliability than traditional white light

sources (such as incandescent lamps, cold cathode fluorescent lamps). Pc-white LEDs must

undergo qualification testing before being released to market. However, most of the

traditional qualification test techniques, such as Failure Modes, Mechanisms, and Effects

Analysis (FMMEA), Fault Tree Analysis (FTA), the Lifetime Test, and the Accelerated Life

Test (ALT), are time-consuming and expensive, especially for devices with long lifetimes.

Sometimes the duration of the reliability test and assessment procedure is longer than the

time between product updates. Therefore, prognostic qualification testing based on historical

test data is desirable for faster commercialization of pc-white LEDs.

From previous studies, it has been found that both lumen depreciation and chromatic state

shift are considered to be the two dominating failures in white LEDs. Most attention has

been paid to the lumen depreciation of LED products, ignoring the chromaticity state shift. In

the LED industry, the Illuminating Engineering Society of North America (IESNA)

(IES-LM-79-08 and IES-LM-80-08) has recommended test methods for measuring the

chromaticity characteristics of LEDs. Additionally, the Next Generation Lighting Industry

Alliance (NGLIA), with the U.S. Department of Energy (DoE), has recommended using

chromaticity shift as an indicator of a white LED’s “end of life”. The American National

Standard Lighting (ANSI) group has also developed specifications for the chromaticity of

solid state lighting products, but they have not demonstrated a chromaticity state shift

prediction method for white LED lighting.

In this work, we use chromatic coordinates (u',v') in the CIE 1976 color space to represent

the chromaticity state of the pc-white LED. The Euclidean distance is the difference between

the original chromaticity coordinates (u'0,v'0) and the future coordinates (u'i,v'i); Δu'v' is

used to describe the chromaticity state shift after aging. Currently, there is no specific

physical model to describe the chromaticity state shift. We present a data-driven prognostic

approach to predict the future chromaticity state based on the observed chromaticity state.

First, we model the chromaticity state shift process by a nonlinear dual-exponential model.

Then we use a recursive nonlinear filtering method (an Unscented Kalman Filter (UKF)) to

predict the future chromaticity state. Finally, we compare the prognostic results to the

extrapolated results. The results show that the Unscented Kalman Filter approach can

improve the prognostic accuracy compared to the conventional extrapolating approach.

Keywords: Phosphor-converted White LED, Chromaticity State Shift, Prognostics, Unscented

Kalman Filter.

47

Project Title: A Flexible Capacitive Micromachined Ultrasonic Transducer

(CMUT) Array

Name of Student: Chong Po Fat, John

Degree: MPhil

Chief Supervisor: Dr C.H. Cheng

Co-supervisor(s): Prof. Y. P. Zheng (BME)


Office: EF 403


A flexible capacitive micromachined ultrasonic transducer (CMUT) array with increased

effective capacitance from the concave bottom electrodes is proposed for ultrasonic imaging.

A CMUT can transmit and receive ultrasound by vibrating its membrane like a drum. DC bias

is applied to bring the membrane closer to the bottom electrode for increasing its sensitivity.

However, most of the developed CMUTs have flat bottom electrodes, which cannot comply

with the deflected membrane in a concave surface. Based on a theoretical analysis, when

using concave bottom electrodes, the effective capacitance can increase 10 times compared

with using flat bottom electrodes.

The current CMUTs developed around the world have several problems when compared with

piezoelectric transducers. First, a CMUT has much lower device capacitance that can be

affected by the parasitic capacitance from the interconnects, wiring, and electronics. Second,

the output pressure is much smaller since the electrostatic force is much lower than the

piezoelectric force, which can cause a lower penetration depth during ultrasonic imaging.

Third, most of the CMUTs are fabricated on silicon substrates that need to be thinned down to

become flexible. However, a thin silicon substrate is very easy to break when standing alone

or when attached to a curved substrate. This project aims to solve these problems in order to

make the CMUT an attractive alternative to the piezoelectric transducer due to its higher

bandwidth, together with the improved device capacitance, output pressure, fill factor, and

sensitivity.

In this study, we will fabricate the CMUT device using bonding technology. The electrode

side with a concave bottom electrode is fabricated by the heat reflow of a photoresist and

nickel electroplating. The membrane side is fabricated by nickel electroplating and it is

bonded to the electrode side by using a photoresist as adhesive. The flexibility is improved by

a novel rivet structure. In combining flexibly with the concave bottom electrode, the

fabricated CMUT can be wrapped around different body parts with different curvatures and

the performance in terms of fill factor, output pressure and sensitivity can be improved.

Keywords: Ultrasonic Transducer, CMUT, Membrane, Flexible array, Micro-machine

48

Project Title: Modelling of Customer Satisfaction and Determination of

Specifications for Product Design Using Computational

Intelligence Techniques

Name of Student: Jiang Huimin

Degree: PhD

Chief Supervisor: Dr C.K. Kwong

Co-supervisor(s): Dr Andrew Ip


Office: EF403


The success of new products is heavily dependent on the associated customer satisfaction

level. If customers are satisfied with a new product, the chance of the product being

successful in the marketplace would be higher. Globalization and an increasing emphasis on

customer needs have dramatically changed the business environment of most industries.

Vigorous challenges have transformed many manufacturers from being

production-centralized to being customer-driven. Therefore, customer satisfaction has

become an important issue that companies need to address while undertaking new product

development.

One of the common ways of developing customer satisfaction models for product design is

based on quality function deployment (QFD). It uses the knowledge, experience, and insight

of product development teams to map customer needs to engineering requirements using a

house of quality. Then, various techniques can be applied to develop customer satisfaction

models. Kansei engineering is a methodology that is commonly used to develop another type

of customer satisfaction model for affective design. Affective design has been shown to

excite the psychological feelings of customers and can help improve customer satisfaction in

terms of emotional aspects. Previous studies have attempted to develop customer satisfaction

models using statistical regression, fuzzy regression, neural networks, quantification analysis

I, and fuzzy rule-based modeling. However, in previous work, explicit customer satisfaction

models that capture both the fuzziness and nonlinearity were not developed. In addition, in

previous studies, the customer satisfaction based on QFD and the affective design were

separated. However, it is quite common that some engineering requirements and design

attributes respectively studied in QFD and affective design have commonalities. Therefore,

the two types of customer satisfaction models need to be considered simultaneously for

determining the optimal settings of the design attributes and engineering requirements of

new products.

In this research, chaos optimization algorithm based fuzzy regression has been developed to

model the relationships between customer satisfaction and the engineering requirements

based on QFD. A particle swarm optimization based adaptive neural fuzzy inference

system has been proposed to generate customer satisfaction models for affective design.

Chaos based NSGA-Ⅱ has been introduced to determine the optimal settings of the

engineering requirements and the design attributes of a new product based on an

optimization model that involves the two types of customer satisfaction models. A case study

of mobile phone design has been conducted to illustrate and evaluate the effectiveness of the

proposed methodology.

Keywords: New Product Development, Customer Satisfaction Model, Quality Function Deployment

(QFD), Affective Design.

49

Project Title: In-Process Visualisation for Deformation Diagnosis in

Hydroforming

Name of Student: Kot Wai Kei Ricky

Degree: MPhil

Chief Supervisor: Dr L.C. Chan

Co-supervisor(s): Dr C.Y. Chan


Contact Number: 27666614

Office: CD101


Hydroforming has been developed over the years, yet it is still considered to be an advanced

technology that has been adopted mainly in the automotive and aero industries. As a result,

this Teaching Company Scheme project is focused on the Hydroforming process at the

request of the collaborating company. Deformation is a critical factor to a successful

Hydroforming process, however, the deformation is wholly concealed in the tooling system

during the hydroforming process, making it difficult for engineers to realize the deformation

of the specimen.

As a result, a visualisation system to analyse the deformation process and interactively

display graphical results in real-time is beneficial but not limited to the above field, enabling

engineers to gather in-process data and see how these parameters affect the deformation

process. Therefore, this study aims to develop an in-process visualisation system for the

diagnosis of deformation. In this study, a proposed visualsation system will be developed to

monitor and to control the entire metalforming process. The in-process deformation

parameters will be collected through a developed Distributive Tracking System, further

processed and then be displayed through an HMI on a display unit. With such a system, the

deformation can then be visualized and be available to engineers for analysis and diagnosis.

The visualized result will be verified by comparing the results with predicted results from

simulation models and the actual product outcomes.

For years, engineers have tried to understand and improve metalforming. However, thus far,

engineers have only been able to understand and analyse the processes before and after the

deformation process. Even with the help of simulation models, one can merely obtain a brief

prediction of the deformation process. Upon the successful development of such a

visualisation system, the real-time deformation status can be obtained, and hence the

situation can be resolved. With this developed in-process visualisation system, engineers will

be able to have a more intuitive understanding of the deformation process and thus enable a

more thorough analysis to be undertaken.

Keywords: Hydroforming, Deformation, HMI, Visualization.

50

Project Title: Failure Analysis of Titanium Tailor-welded Blanks under

Multi-stage Forming Process

Name of Student: Lai Chi Ping

Degree: PhD

Chief Supervisor: Dr. L. C. Chan

Co-supervisor(s): Prof. T. C. Lee

Prof. C. L. Chow (University of

Michigan-Dearborn)


Office: CD101


Production of products that are light-in-weight and small-in-size is becoming the trend in

modern industry. Manufacturers are looking for new materials or methods for production.

One of the methods is to use the lighter materials for production, instead of traditional steel,

such as light-weight alloys. Different light-weight alloys have their own characteristics, high

working temperature, high strength and high corrosion.

Titanium alloy is not widely used in the automotive industry because of its high cost, but

there is a great potential for the use of titanium alloys in the aircraft, space, marine and

military industries. However, due to the relatively low ductility at room temperature and high

yield strength, Ti-TWBs cannot be formed successfully by using traditional forming

methods. Failures are usually caused by the wrong setting of the forming parameters.

The multi-stage forming process entails a series of operations which converts sheet metal or

TWBs into a part of the desired shape without any fractures. However, engineers often

encounter the problem of poor formability for Ti-TWBs and otherwise desirable lightweight

alloys, which have impeded their widespread utilization in the automotive and aircraft

industries. A problem arises from the failure of the sheet during the multi-stage forming

process, typically as necking, due to strain localizations that develop in the material. In fact,

some researchers have investigated the formability of sheet metal influencing the change of

strain path and material properties by means of finite element simulations and experiments,

but not for Ti-TWBs. Therefore, many manufacturers are looking forward to being able to

form Ti-TWBs in more effective ways.

In this study, a new damage-based failure prediction model, involving a mixed

isotropic-kinematic hardening rule and anisotropic damage mechanics, for the multi-stage

and hot forming operations of Ti-TWBs, will be developed with the aid of experimental

validation and material analyses. The damage-based failure model will be developed to

examine the deformation behavior for Ti-TWBs under multi-stage forming processes. This

workable model will assist engineers to predict and virtually optimize the main technological

parameters of the Ti-TWBs forming process before its physical realization.

Keywords: Titanium, Tailor-welded blanks, failure analysis, multi-stage, forming


51

Project Title: Design of a New Molding Process for Making Seamless Hollow

Plastic Parts

Name of Student: Ng Wai On

Degree: MPhil

Chief Supervisor: Dr C.Y. Chan

Co-supervisor(s): Prof. K.L. Yung, Dr H.Y. Chen


Office: EF403


This is a teaching company scheme project funded by the Innovation and Technology Fund

from the Hong Kong Government, and is jointly organized by the Department of Industrial

and Systems Engineering (The Hong Kong Polytechnic University) and the Mattel (HK)

Limited.

Many seamless hollow plastic parts are fabricated by rotational molding due to the

advantages of low pressure characteristics, simple tooling, and uniform thicknesses.

However, the process is quite labour intensive. Besides, the working condition is usually

poor and the energy efficiency is low. Thus, there is a potential need for a better process as

the demand is huge, especially, in the toys industry. The new process should consider both

the cost effectiveness and product quality as well.

The imminent application of this process in Mattel is for the production of the Barbie dolls

that can reach 30 million units per year in Mattel’s China plant. With the increasing labour

cost in China, automation is also an important issue in the design.

Our current research focus is on the utilization of the induction heating method to replace the

current diesel heating process used. This is because the use of diesel ovens is a main cause of

having poor working conditions and low energy efficiency. It is believed that a considerable

improvement can be achieved by the application of electrical heating in the process. The

success of this project will benefit manufacturers who are using rotational moulding for

producing products.

Keywords: Process Heating, Plastic, Hollow, Seamless, Manufacturing.

52

Project Title: A New Production Model to Compensate Forecast Error and

Customer Loss in Waiting

Name of Student: Chen Qian

Degree: PhD


Co-supervisor(s): Prof. K.L. Yung


Office: EF403


Production systems play an important role in modern society, and significant developments

have been achieved over the years. However, attributable to the diversity of human behavior,

customer demand uncertainty always exists in practice. Consequently, a production solution

that is capable of handling such an unforeseen fluctuation is desired. To address the demand

uncertainty problem, forecasting is a typical solution in production management. Thus, there

is always a need for research in improving the forecast accuracy and developing of new

methods.

Apart from working on the new forecast method or improving existing models, this research

is about how to work with the expected forecast error in a most economical way. To achieve

this goal, we aim at making a balance between the effect of forecast error and customer loss

in waiting for the “product”, as time is the key factor that affects the forecast error. In this

research, a production approach named Make-to-Balance (MTB) was developed. To verify

the concept and the operating result of the proposed model, a software program was coded

and showed that an optimal production solution could be obtained from both the program and

the proposed mathematical MTB model. Indeed, the contributions of this research are not

only its inspiration but also extends the view on how to run a production system effectively

by taking into account both the concern with forecast error and customer behavior. It also

signifies that customer loyalty helps to reduce the effect of forecast error.

Keywords: production model, forecast variation, customer behavior

53

Project Title: A Flexible 2D Piezoresistive Shear and Normal Force Sensor

Array for Pressure Mapping Applications

Name of Student: Shi XiaoMei, Sissi

Degree: PhD

Chief Supervisor: Dr C.H. Cheng

Co-supervisor(s): Prof. Alex Wai (EIE)


Office: EF403


It is important to identify the human body 3D reaction force produced in the study of

human-machine interaction and biomechanical analysis, and some sensor systems have been

developed for applications in human dynamics analysis. In the market, most of the force

sensors only measure the magnitude of the resultant force. Capacitive normal and shear force

sensors with polymer substrates have been reported. They use parallel plate capacitors with

four electrodes to differentiate normal and shear forces. A normal and shear piezoresistive

tactile sensor that uses inclined micro-cantilevers covered with an elastomer has also been

developed with crystalline silicon as the piezoresistive material. All of them use

polydimethylsiloxane (PDMS) as a flexible substrate material to encapsulate the sensing

materials. However, solid conductors are still used as sensing materials, and can be broken

when a large force is applied.

This project aims to fabricate a flexible normal and shear force sensor by using liquid metal

alloys (Ga-In-Sn) as piezoresistors, to improve the durability of the sensing material. By using

liquid metal alloy as the gauge material, it can detect large forces without breaking the sensor

wires. Since the liquid-metal piezoresistors deform with the elastomeric substrate, shear and

normal forces can be detected with resistance changes of the piezoresistors. Each force sensor

comprises a pair of symmetric piezoresistors, which are screen-printed on the cavity of the

PDMS substrate with a tilt angle around 30° so as to be sensitive to both normal and shear

forces. The serpentine design is to reduce the self-heating effect by increasing the resistance

of the metallic gauge wire. The initial resistance of the strain gauge was 10 ohm on average.

The force can be measured with the strain-induced resistance change that depends on the

gauge factor (GF). Normal force will compress both piezoresistors as a common mode while

shear force will shorten one piezoresistor but elongate the other as a differential mode. The

test results demonstrate the sensitivity of the force sensor in both the normal and shear

directions. The hysteresis of the force sensor was also measured. The response time of the

strain gauge depends on the speed of the applied force due to the PDMS elastomeric substrate.

Keywords: normal and shear force sensor, PDMS, liquid metal, styling

54

Project Title: A System Monitoring Model by Examining Entity Dynamics

Name of Student: Wang Lei

Degree: PhD




Office: EF 403


Monitoring plays an essential role in manufacturing processes as it serves as an information

bridge between the system status and the necessary correction activities. Generally, for a

Computer Integrated Manufacturing (CIM) system, the monitoring procedure comprises

collecting various signals from the facilities and building up a dedicated mathematical

model to interpret these signals. The process is analogous to a doctor who diagnoses a

patient by checking organs individually. However, from the managerial viewpoint, the

healthiness of the whole system is more needed than the operational detail. Inspired by the

distinctive philosophy that a proper system should be working in harmony in sense, a novel

method for presenting a holistic picture of a manufacturing system by examining the flow

entities is presented in this research.

This research has three stages. First, a manufacturing system is modelled as the integration

of a set of Region Of Interests (ROIs) in a top-down manner. Second, analogous to the

concept of checking blood pulses, several features are extracted from a system to constitute

the “pulses” of an ROI, and these include the Regional Inconsistency (RI), the

Inter-component Arrival Time (IAT), the Inter-component Leaving Time (ILT), and the

Instant Work-In-Progress (IWIP). Then, a reasoning scheme is determined to detect two

types of popular abnormalities, blockages and slowdowns, in an ROI. Third, an ROI

segmentation technique is developed to assist the monitoring framework design by taking

into consideration the tolerable system response time.

It is anticipated that based on analyzing the “pulses” tones of all ROIs, the healthiness of

the holistic system can be reflected, and simulation experiments will be conducted to

validate the effectiveness of the proposed monitoring approach in this research. In terms of

the hardware requirement, only simple counter devices with time-stamp functions are

needed, and this highly enhances the portability of this proposed approach.

Keywords: production monitoring, material handling, counter deployment

55

Project Title: A Novel Metaheuristic Model with Distributed Pattern

Learning

Name of Student: Xue Fan

Degree: PhD


Co-supervisor(s): Dr Andrew Ip, Prof. Benny Cheung


Office: DE406


Automated heuristic selection and heuristic generation have increasingly attracted attention

in solving combinatorial optimization problems emerging from both theory and practice.

This research presents a heuristic generation algorithm, called Suboptimum- and

Proportion-based On-the-fly Training (SPOT), which can enhance existing heuristics with

the aid of instance-specific information.

By making use of the proposed “sample-learn-generate” framework, SPOT initially

samples small-scale subproblems. Then, it collects the instance-specific information from

the suboptima of the subproblems by means of machine learning. Lastly, it generates new

heuristics by modifying existing heuristics and data structures.

In the development of SPOT, two standards were incorporated to regulate the problem input

and the machine learning data. The software implementation was done in Java, with two

external development libraries, the HyFlex and the Weka. In terms of testing, two

well-known NP-Complete combinatorial optimization problem domains were employed:

the Traveling Salesman Problem (TSP) and the permutation Flow-Shop scheduling Problem

(FSP).

Each generated heuristic was tested with the TSP and the FSP domains. To verify the result

of using SPOT, one of the winners of the international hyper-heuristic competition CHeSC

2011, named PHunter, was tested with the generated heuristics by SPOT. In the TSP,

adapting SPOT gave little improvement, but in FSP, the improvements were significant. It

increased the overall score of the PHunter from 20.5 to 43 (out of 50). Indeed, it also

outperformed the best records in CHeSC 2011: 32 by AdaptHH, 29.5 by MLand 26 by

VNS-TW.

Keywords: Hyper-heuristics, automated algorithm generation, sampling-based instance-specific

algorithm, Suboptimum- and Proportion-based On-the-fly Training

56

About us

Our graduate institutions before being an ISE research student

Our Hobbies / Interests

57

Our characteristics

Our future plans

58

Research Students' Whereabouts

D

E406

59

DE

404

60

E

F403

proceedings ofthe fabrication of different types of 3d-structured surfaces (i.e. aspherical,...

Documents