RB 24

CIVIL ENGINEERING
RESEARCH ISSN 0219-0370 • No. 24 / 2011 • www.ntu.edu.sg/cee/research/bulletin/index.asp
SUSTAINABLE URBAN SYSTEMS

In 2000, the world population was 6.1 billion. The of waste that must be handled safely without impacting the
Population Division of the United Nations projected that quality of modern urban living.
this would grow to 8.9 billion by 2050, an increase of
nearly 47 per cent in 50 years. Dr Su Guaning, President NTU, in his opening remark at
the GlobalTech’s Workshop on Sustainable Urban Solutions,
Today, half the world’s population lives in urban areas. held at Shanghai Jiaotong University, China on 03-May-
By 2050, eight out of ten people on planet earth will be 2010 said:
living in cities. This prompted the Executive Secretary
of the International Convention on Biological Diversity “For a long time, mankind has made fundamental
to say that “the battle for life on Earth will be won or assumptions on Mother Nature that turn out to be grossly
lost in cities”. inaccurate and highly dangerous. From a small population
dependent largely on processes of nature for our survival,
With the explosive growth in population and the unrelenting we have evolved and multiplied and acquired the powers
shift to urbanization, successfully engineering and building to transform nature for the good and the bad. No longer
sustainable cities within limits imposed by the natural can we take the impact of mankind as negligible and
environment is a huge challenge. Cities consume two-thirds the capacity of nature as infinite.” He commented that
of total energy produced and generated over two-thirds “Unfortunately we are in the awkward situation where we
of global energy-related CO2 emissions. Cities impose
tremendous demand on natural resources such as clean water
and clean air. In their wakes, they created a huge volume
Cont’d on Pg 3
MESSAGE FROM THE CHAIR

Dear friends: and in building energy efficiency. Both are somewhat
non-traditional areas for the School and representative
O ur School of Civil and

Environmental Engineering (CEE)
has continued its pace of rapid growth
of NTU’s renewed efforts in intensifying integrative
research as well as being aligned with NTU’s strategic
peak of excellence in Sustainable Earth. The first is the
over this past year strengthening its establishment of NTU’s Institute of Catastrophe Risk
research programmes in the sustainable Management (ICRM) in January 2010 with CEE being
built environment domain. Early in the key driving School and CEE Professor T-C Pan,
2010, CEE accounted for two of NTU’s four prestigious Dean of College of Engineering, being the Founding
awards in Singapore’s National Research Foundation
Director. ICRM aims to become Asia’s leading research
(NRF) Competitive Research Programme call on
Sustainable Urban Systems. These extremely competitive institute in catastrophe risk management, helping those
awards of about S$10 million each saw faculty teams at risk worldwide in general and Asia in particular. The
across CEE’s three Divisions led by A/P Chu Jian and focus on Asia is because Asia suffers from having the
A/P Wang Jing-Yuan winning awards in the development largest exposure and corresponding fatalities and losses
of an underwater city and in sustainable urban waste from recent major catastrophes but is the least prepared
management, respectively. The feature article of this issue from both societal and governmental fronts. Working
of CEE Research Bulletin contains detailed descriptions synergistically with other NTU colleges and institutions,
of both research programme which are ramping up. ICRM will undertake integrated risk assessment and
This year also saw CEE embarking on two new research

fronts, that specifically in catastrophe risk management Cont’d on Pg 2
IN FOCUS
Message from the Chair (Cont’d from Pg 1)

management of hazards and catastrophes, both natural and research activities being consolidated to free up spaces
man-made. As a start, ICRM is developing methodologies for research offices and laboratories. In particular, CEE
for the rigorous assessment of seismic risk and flood now has a new laboratory, the Central Environmental
risk. These two efforts are led by CEE Emeritus and Science Engineering Laboratory, housing several
Professor C.N. Chen and CEE Professors Shuy E.B., Li new instruments for environmental research and a new
Bing and Kusno Megawati. Together with the Nanyang Civil Engineering Materials Laboratory for green concrete
Business School, ICRM has also secured PhD funding research. I encourage you, and particularly the alumni,
for two students from Aon-Benfield, a global leading to visit and see the transformations first hand.
reinsurance intermediary. One of the PhD students will
focus on catastrophes in maritime business and the other A key part of School management is to ensure that CEE’s
on Asian motor liability. Additional post-doctoral fellow education and research programmes are of the highest
funding is also secured from Willis Re, another global possible level in rigor and quality. Notably, CEE saw
reinsurance firm, for research into marine container and visits by two key visiting committees this past year.
cargo exposure/vulnerability, and for CEE to join the The first is by NTU’s College of Engineering Visiting
Willis-Re Global Research Network. CEE Professors Committee and the second is by Singapore’s Ministry of
Robert Tiong, Jasmine Lam and Chiew Sing Ping are Education External Review Panel under the Ministry’s
leading these efforts with the Aon-Benfield and Willis- Quality Assurance for Universities (QAFU) framework.
Re grants. I am happy to report that CEE has received positive
comments and reviews by both committees. Furthermore,
The second research front is in energy efficiency for our Bachelor of Engineering degree programmes formerly
buildings in tropical climates. Buildings in Singapore accredited by UK’s Joint Board of Moderators (formed
account for 50% of the nation’s energy usage (the number by The Institution of Civil Engineers, the Institution of
is about 40% in the US) representing the nation’s largest Structural Engineers, the Institution of Highways and
energy demand. NTU, led by CEE and its School of Transportation, and the Institute of Highway Incorporated
Electrical and Electronic Engineering (EEE), partnered Engineers), are now accredited by the Engineering
UC Berkeley thereby enabling Berkeley to establish its Accreditation Board of the Institution of Engineers
Berkeley Education Alliance for Research in Singapore Singapore under the Washington Accord. At the individual
(BEARS) under NRF’s CREATE scheme. BEARS will level, Dr Goh Kok Hui, a recent PhD graduate under the
have the Building Efficiency and Sustainability in the supervision of Professor Lim Teik Thye, won one of the
Tropics (BEST) programme as its first programme. Under World Future Foundation’s PhD Prize in Environmental
BEST we will see Berkeley faculty partnering CEE and and Sustainability Research for his thesis on the sorption
EEE counterparts to develop and testbed technologies that of oxyanions on nanocrystalline Mg/Al layered double
will significantly reduce the building energy demand. The hydroxides. CEE Professor T-C Pan, Dean of College of
focus will be on self-optimizing building systems using a Engineering, won the 2010 Defense Technology Prize for
combination of physics-based models, models of occupant work over the years in underground technology and rock
behavior, data driven models and smart sensing to deliver engineering, and thereby provided key support towards
custom environments that simultaneously optimizes the current developments of underground facilities in
comfort and productivity, and minimizes energy and Singapore.
environmental costs. The research teams aim to achieve
disruptive innovations in building energy systems which CEE continues to be successful in attracting new,
will be vital in creating new knowledge for new buildings high calibre faculty members over this past year.
and retrofitting of existing buildings. Professors Chiew Since January 2010 we welcomed 5 new faculty
Sing Ping, Susanto Teng and Victor Chang and I will members. They are Associate Professor Tarek Zayed
be heading CEE efforts on this front. (Construction Engineering Management, PhD Purdue
University, 2001) Assistant Professors Lee Chang Soo
It is indeed most exciting to see these new large research (Environmental Engineering, PhD Pohang University of
programmes developing over the course of the year. Science & Technology, 2005), Yang En-Hua (Civil and
Furthermore, these initiatives are coming on top of the Environmental Engineering, PhD University of Michigan,
Civil Engineering Research • January 2011
research centres and programmes started under NTU’s 2008), Philip Wong (Environmental Engineering and
Nanyang Environment and Water Research Institute Science, PhD Stanford University, 2010), Cheung Sai
(NEWRI) with again CEE being the lead School and Hung (Civil Engineering, PhD California Institute of
which I have reported on in last year’s issue of CEE Technology, 2009).
Research Bulletin. On the flip side, these many new major
initiatives are placing dramatically increased demands I hope you find CEE Research Bulletin informative and
on CEE’s research infrastructure to the point that CEE enjoyable. Please do e-mail me with your thoughts and
is now facing major shortages in office and laboratory comments.
space at its N1-Block on campus. CEE currently has
over 240 PhD students and over 120 research staff
comprising research fellows, research associates and
project officers. We have thus embarked on several Edmond Lo
major renovation projects that will see teaching and Chair, CEE
2
IN FOCUS
Sustainable Urban Systems (Cont’d from Pg 1)

need to know the impact of our actions on nature without city can be constructed using the approach proposed. Living
sufficient understanding of nature itself.” under seawater will be a reality in the near future.
He further announced at the workshop that, “a happy The proposed new approach has many advantages over
coincidence of our prior preparation, the increase of the existing practices. It makes multiple use of sea space
external funding as well as our convictions has resulted by making the space both above and below the reclaimed
in an explosion of energy and activities in the Sustainable land available for recreation, living or infrastructural
Earth Peak of Excellence of our NTU 2015 Strategy. It is development. It combines reclamation, superstructure and
backed by a multi-year funding of S$700mil on a diverse underground constructions in one and thus is the most
range of topics ranging from earthquake and volcanoes to efficient and cost-effective approach for space creation
environmental engineering to biofilms to energy to water and utilization. When designed strategically, the cylindrical
resources to membrane technology.” structure groups can also function as effective shore
protection elements against extreme waves such as storm
The School of Civil and Environmental Engineering is a surges or tsunami and seawater changes caused by global
bulwark in the preparation and on-going effort to spearhead warming. They can also be designed to create energy
the Sustainable Earth Peak of Excellence in NTU 2015 using waves.
strategy. The School had won two Competitive Research
Programme (CRP) funded by the National Research The design and construction of the super large cylindrical
Foundation. structures poses many challenges to both theories and
practice. These include (a) the development of new and
The first project titled “underwater infrastructure and innovative construction materials that would allow the
underwater city of the future” is focused on the exploration massive concrete structures to be constructed economically
of new approaches to land-use particularly in the near- and last for hundreds of years; (b) innovative soil
shore environment. This research effort is particularly improvement and foundation construction methods to
timely and pertinent to land scarce Singapore where other allow a proper control and prediction of the settlement
potential approaches to expanding land-space are severely and bearing capacity under various loads; and (c) the
constrained. hydrodynamics and related coastal, environmental issues
The second project is titled “Sustainable Urban Waste

Management for 2020”. The main focus of the project is
directed at the decentralized “waste-to-resources’ concept,
targeting at the recycling, reclamation and reuse of all
forms of wastes generated in highly urbanized cities of
which Singapore is a prime test field.
The following provides a synopsis of the two CRPs.
CRP I:
UNDERWATER INFRASTRUCTURE AND
UNDERWATER CITY OF THE FUTURE
Figure 1. Use of large size cylindrical structures for space
In land scarce Singapore, space creation is a key strategic creation underwater and for land reclamation above.
area that concerns the survivability and sustainability of the
Nation. At the present, we have been using underground
caverns and offshore reclamation for space creation. Both

methods may not be sustainable in the long run. A new
approach – going underwater is proposed to make use of
the sea space to construct underwater infrastructure and
at the same time use the top-side of the infrastructures
as reclaimed land (see Figure 1). Using the proposed
method, we no longer need to import fill materials for land
reclamation. The proposed method also allows the use of
our limited space even more efficiently and cost effectively.
Cylindrical structures can also be put together to form a
watertight seawall and thus create space behind the seawall Figure 2. Use of large diameter cylindrical structures to form
(see Figure. 2). What is more exciting is that underwater seawalls and create space behind the seawall.
3
IN FOCUS
and the possibility for harvesting wave energy. New CRP II:
technology and innovative solutions are required to make SUSTAINABLE URBAN WASTE MANAGEMENT
the idea a reality. It is proposed in this project to carry out FOR 2020
an intensive inter-disciplinary study to address problems
INTRODUCTION
related to structural, geotechnical, hydrodynamic, risk
analysis, and socio-technical aspects. From the study, a Waste is not waste. Much of our disposed waste should
series of innovative methods, new materials, and new not in fact be considered waste; they are simply misplaced
construction technologies will be developed to make the resources. Based on current waste management concepts,
proposed approach technically feasible and cost-effective. these resources are normally buried in landfills or incinerated.
The scopes of the study include (a) in-depth study of the Such waste treatment/disposal approaches need to be revised
mechanical properties and durability of new construction as natural resources are depleted. Waste becomes potential
materials including Ultra High Performance Concrete, sources for resource recovery. This is especially true for
land- and resource-scarce Singapore.
Self-Compacting Green Concrete, as well as innovative
cylindrical structural design using new forms of structural A successful proposal entitled “Sustainable Urban Waste
members; (b) Innovative soil treatment and foundation Management for 2020”, based on a decentralised “waste
methods; and (c) hydrodynamic study and harvest of wave to resources” concept, was recently awarded S$10 million
energy. The innovative methods to be developed include by the National Research Foundation (NRF) under its
the use of green concrete and new reinforcement, new Competitive Research Programme (CRP) funding scheme
methods for the installation of cylindrical structures such to develop sustainable urban waste management solutions
as suction caissons without soil improvement, new methods for 2020 and beyond. The five-year programme is led by
Assoc Prof Wang J.Y. (lead PI) of the School of Civil and
to address issues on vortex shedding, wave trapping and
Environmental Engineering (CEE) and supported by six
movement of sand waves relevant to the founding of extra other NTU faculty members as Co-PIs. The programme
large structures on the seabed, and methods to harness also partner with three government agencies, two industrial
wave energy. companies, and two overseas universities.
The expected outcome of this research is the establishment

of a new, efficient and sustainable method for space creation OBJECTIVES
using less resource and to provide innovative solutions
The main objective of this research programme is to
to the challenges to make space creation and underwater develop sustainable urban waste management solutions for
construction more cost-effective and socially attractive. 2020. Research outcomes are expected to bring long term
environmental, economical and social benefits to Singapore
and eventually the rest of the world. To achieve the main
KEY MEMBERS OF PROGRAM TEAM: objective, the following three research subprogramme were
proposed:
Principal Investigator: • Subprogramme A: Communities as renewable resource
recovery centres
• Assoc Prof Chu Jian (CEE, NTU)
• Subprogramme B: Wastewater treatment plants as urban
eco power stations
Co-Principal Investigators:
• Subprogramme C: Rapid land reclamation of closed
• Assoc Prof Susanto Teng (CEE, NTU) dumping grounds
• Assoc Prof Tan Soon Keat (CEE, NTU)
Collaborators: RESEARCH OUTCOME

• Assoc Prof Robert Tiong (CEE, NTU) Various technologies will be developed, tested, and
• Asst Prof Sulfikar Amir (AHSS, NTU); demonstrated for all the subprogrammes.
• Mr Lam Kok Pang (JTC Corporation)
Subprogramme A
Industrial collaborator:
• Surbana International Consultants
Overseas Collaborators:
• Mr Knut H. Andersen, Norwegian Geotechnical
Institute, Norway;
• Dr Dale Berner, Ben C. Gerwick Inc., USA;
• Prof Odd E. Gjorv, Norwegian University of Science
and Technology, Norway;
• Prof Richard S. Mercier, Offshore Technology
Research Centre, Texas A&M University, USA;
• Prof Andrew Whittle, MIT, USA
4
IN FOCUS
• A technology for phosphorous - and nitrogen-recovery KEY MEMBERS OF PROGRAM TEAM:

from yellow water
• A technology for odour removal from household source Principal Investigator:
separation system • Assoc Prof J-Y Wang (CEE, NTU)
• A two-phase anaerobic digestion (AD) system for co-
digestion of black water and food waste Local Collaborators:
• An integrated thermophilic co-digestion (CoD) and co- • Prof W.J. Ng (CEE, NTU): PI on co-digestion of
composting (CoC) system sludge and community organic waste
• An integrated solid oxide fuel cell (SOFC) + microbial • Prof R. Harianto (CEE, NTU): Co-PI on capillary
electrolysis cell (MEC) + proton exchange membrane capping development using recycled materials
fuel cell (PEMFC) system • Assoc Prof X. Wang (CBE, NTU): Co-PI on
integrated SOFC + MEC + PEMFC system
Subprogramme B • Assoc Prof J. Chu: Co-PI on non-invasive site
characterization technique development
• Asst Prof Victor Chang: Co-PI on nutrient recovery
and odor removal
• Asst Prof C.S. Lee: Co-PI on co-digestion of black
water and food waste
• Mr S.H. Lim (National Environment Agency):
Partner on the whole programme
• Ms Cherlyn Leong (Jurong Town Corporation):
Partner on subprogrammes A and C
• Ms P.S. Teh (Housing Development Board): Partner
on subprogramme A
• Mr William Phay (Keppel Pte Ltd): Partner on
• An efficient sludge + organic waste co-digestion (single subprogramme B
stage + thermophilic ) system
• Mr S.O. Goh (SembCorp Pte Ltd): Partner on
• Other developments: A two-phase AD for sludge + subprogramme C
organic waste co-digestion
Overseas Collaborators:
Subprogramme C
• Prof Rainer Stegmann (Technical University of Hamburg
and Harburg, Germany): Collaborator on integrated co-
digestion and co-composting system
• Prof J.S. Chang (National Cheng-Kung University
(NCKU), Taiwan): Collaborator on co-digestion of black
water and food waste and landfill remediation
• Prof S.S. Cheng (NCKU, Taiwan): Collaborator on
bioenergy development
• A technology for non-invasive geophysical site

investigation
• Landfill mining technologies including waste
characterization and separation
• Remediation technologies including electrokinetic, soil
washing, bioremediation, etc.
• Resource recovery technology, e.g., fast pyrolysis
• Evapotranspiration-based (capillary) capping technology
5
CEE VISION AND MISSION
CEE VISION & MISSION
Our Vision
A leading school for sustainable built
environment.
Our Mission
To nurture students to be responsible leaders
capable of realising their maximum potential
in their profession and community. To provide
a collegiate environment for faculty to excel
in education and research for sustainable
development. To advance knowledge for the
practice of civil and environmental engineering

and maritime professions.
6
STATISTICS
STATISTICS
Faculty & Staff (as of 1 December 2010)
Publications, Patents and Research Grant
Year 2006 2007 2008 2009 2010

Journal 145 167 151 202 79
papers
Conference 76 120 91 90 45
papers
Patents 1 3 3 2 2
Research 9.6 9.9 13.5 36.4 30.0*
Grant ($mil)
*Partial figure only
Students Enrolment
Programme/ Under- MEng PhD PhD MSc MSc MSc MSc

Academic Year graduate (SSP) (Civil Eng) (Env Eng) (ICM) (MS)
AY2006-07 1364 36 145 3 254 102 79 175
AY2007-08 1290 37 150 6 228 107 102 201
AY2008-09 1210 29 163 6 216 93 98 256

AY2009-10 1148 5 169 7 38 14 25 22
AY2010-11* 1149 5 206 7 54 31 45 49
*Semester 1 only
7
UNDERGRADUATE PROGRAMMES
BACHELOR DEGREE PROGRAMMES

BACHELOR OF ENGINEERING (CIVIL) BACHELOR OF ENGINEERING
(ENVIRONMENTAL)
The Civil Engineering programme is structured on a flexible
and diverse system that allows students to choose from a The Environmental Engineering programme is structured on
broad range of courses to receive a well-rounded education a flexible and diverse system that allows students to choose
while maintaining high academic standards. In the final year from a broad range of courses to receive a well-rounded
of study, students are given the opportunity to specialize in education while maintaining high academic standards. In
particular areas of civil engineering by selecting the relevant the final year of study, students are given the opportunity to
elective courses. Furthermore, though students typically specialize in particular areas of environmental engineering
complete the degree course in four years, they may study by selecting the relevant elective courses. Furthermore,
at their own pace and complete their studies within the though students typically complete the degree course in
time frame of three-and-half to seven years. four years, they may study at their own pace and complete
their studies within the time frame of three-and-half to
In the first year of study, students take the common seven years.
engineering course which deals with basic concepts
in mathematics, science and fundamental engineering In the first year of study, students take the common
principles. The curriculum also includes a course in engineering course which deals with basic concepts
communication skills and two courses of laboratory in mathematics, science and fundamental engineering
experiments. principles. The curriculum also includes a course in
communication skills and two courses of laboratory
In the second year, students are required to take fundamental experiments.
courses in the civil engineering discipline, such as basic
theory of structures, geotechnical engineering, water In the second year, students are required to take fundamental
resources engineering, engineering drawing & measurement. courses in the environmental engineering discipline, such
Second-year students also take additional courses in as fluid mechanics, hydrology, environmental chemistry,
mathematics, two courses of laboratory experiments and environmental processes, and environmental microbiology.
a technical communication skills course. In the second Students are given foundational training in sustainable
semester of the second year, students study “Engineering infrastructure by taking some courses in basic theory of
Innovation and Design”, a course in which students work structures and engineering drawing & measurement. Second-
in groups on a given open-ended project to learn the year students also take additional courses in mathematics,
practical and innovative problem-solving skills required two courses of laboratory experiments and a technical
of engineers. communication skills course. In the second semester of
the second year, students study “Engineering Innovation
In the third year of their studies, students are offered a and Design”, a course in which students work in groups
balanced mix of core courses comprising structural analysis, on a given open-ended project to learn the practical and
design in concrete and steel structures, and specialized innovative problem-solving skills required of engineers.
courses in foundation, transportation and environmental In the third year of their studies, students are offered
engineering. After students attaining the requisite academic a balanced mix of core courses comprising water
units for promotion to third year, the students can, if they supply engineering, wastewater engineering, solid waste
wish, register for a 22-week Industrial Attachment (IA) engineering, geo-environmental engineering, hydraulics and
in a private company or public organization, where they basic structural design. After attaining the requisite academic
learn to practise civil engineering under the guidance of units for promotion to third year, the students can, if they
experienced engineers and managers. wish, register for a 22-week Industrial Attachment (IA)
in a private company or public organization, where they

In the final year, the Civil Engineering programme learn to practise the environmental engineering under the
concentrates on training students in professional civil guidance of experienced engineers and managers.
engineering practice as well as managerial and entrepreneurial
skills. Students are given the choice to pursue their own In the final year, the Environmental Engineering
fields of interest in a particular area of specialization by programme concentrates on training students in professional
selecting from a group of optional elective courses offered environmental engineering practice as well as managerial
by the School. Each student is also required to complete and entrepreneurial skills. Students are given the choice to
a two-semester duration research project in any of the pursue their own fields of interest in a particular area of
specializations in civil engineering. In Integrated Design, specialization by selecting from a group of optional elective
students will be involved in team effort to confront and solve courses offered by the School. Each student is also required
real-life open-ended civil and environmental engineering to complete a two-semester duration research project in any
problems. of the specializations in environmental engineering.
8
UNDERGRADUATE PROGRAMMES
BACHELOR OF SCIENCE (MARITIME well versed with the maritime industry and strong business
STUDIES) knowledge so that they will be the future business leaders
in the global maritime industry. MPA and the industry are
The Maritime Studies programme focuses primarily on fully supportive of the Maritime Studies degree programmes
tertiary education in shipping, business, management, and and MPA also provides an endowed Professorship (Shipping
maritime science and technology, to build up the expertise Management) in NTU.
of the local shipping industry as well as working towards
establishing Singapore as a centre of excellence for shipping The BSc (Maritime Studies) curriculum aims to provide
business, research and development. The programme is students with both depth and breadth. The course structure
conducted jointly by NTU and the Norwegian School of is flexible and broad base. Students will be required to
Management (BI), Norway, supported by the Maritime and complete:
Port Authority of Singapore (MPA). • Foundation courses including mathematics, social
sciences, business and technology
With the support from the College of Engineering, • Shipping specialist courses including organization
Nanyang Business School and School of Humanities and management of shipping companies, international
and Social Sciences, students enrolled in the Maritime shipping logistics, maritime law, marine insurance,
Studies programme will learn from academics from various shipping strategy, and a research project
disciplines, thereby developing different skills in a holistic • Prescribed electives for specialisation in the
and comprehensive learning environment. The Norwegian programme, and General Education Requirement
School of Management (BI) is Norway’s second largest courses for broadening the learning experience
educational institution, and one of the largest business
schools in Europe. BI is the first Norwegian educational In addition to the above, the more rigorous BSc (Maritime
and research institution to achieve international accreditation Studies) with Business Major curriculum includes core
establishing BI as one of Europe’s leading business business courses in accounting, business law, company
schools. law, principles of taxation, business environment, financial
analysis & reporting, marketing, and organisation behaviour
The Maritime and Port Authority of Singapore (MPA) and & design.
the shipping industry have recognised that the shipping Students will complete a semester of their studies at BI,
practice and business in Singapore need to be further Norway, in their third year of studies. The curriculum
elevated in order to enter into the regional and global also includes an Industrial Immersion - ten weeks for BSc
arenas. The BSc (Maritime Studies) degree is a strategic (Maritime Studies) and twelve weeks for BSc (Maritime
development that would provide high-level and high-value Studies) with Business Major - where students will be
education for professionals in shipping and related business, attached to organizations in the shipping and related
elevating them from local business management to one of industry.
international business standing. The BSc (Maritime Studies)
with Business Major degree aims to produce graduates
9
GRADUATE PROGRAMMES
GRADUATE STUDIES
BY COURSEWORK Master of Science (Infrastructure Engineering
and Management)
Master of Science (Civil Engineering)
The programme is a joint Degree Programme between the
The programme equips students with the latest advancements School of Civil & Environmental Engineering, Nanyang
in knowledge and technology in modern civil engineering Technological University, Singapore and the Department
practice. Students will also have the opportunity to acquire of Civil Engineering, Indian Institute of Technology
knowledge in several civil engineering disciplines by Bombay, India. The programme provides holistic training
selecting appropriate courses. in infrastructure engineering and management covering
conceptual and physical planning, design, and operational
Master of Science (Environmental aspects of infrastructure systems. Such systems are in great
Engineering) demand in rapidly developing regions such as in China,
The programme equips graduate engineers and other related India, ASEAN and the Middle East and include air, sea
professionals with advanced skills and expertise in a wide and land transport networks, water supply and wastewater
variety of environmental disciplines. The programme systems and power distribution networks.
offers a comprehensive range of subjects on advanced
water and wastewater treatment, air and land pollution
as well as broader aspects of environmental management RESEARCH
and planning. Students can choose to pursue Doctor of Philosophy degree
in one of the following disciplines:
Master of Science (International Construction Construction Technology and Management
Management) Construction Technology and Management
The programme enables graduate engineers, architects and Risk and Project Financing
other related professionals to expand their decision-making Facility and Infrastructure Management
horizons given the kind of parameters and risks which Structures and Mechanics
international construction managers encounter. The main Computational Mechanics
objective of the programme is to develop competent and Earthquake Engineering and Structural Dynamics
well rounded construction managers who have the skills Protective Technology
to source, secure and effectively manage projects in the Fire Engineering
domestic and international construction market. Concrete and Steel Technology
Geotechnical Engineering
Master of Science (Maritime Studies) Foundations of High-Rise Buildings
The programme provides graduate level and high-value Land Reclamation
education for professionals in shipping and related business; Underground Space Development
elevating them from local business management to one of Tropical Soil Engineering
the international and global business settings. The foremost Rock Mechanics and Engineering Geology
intention is for young graduates and middle-management Environmental and Water Resources
executives working in shipping related areas an avenue Membrane Technology in Environmental Engineering
for higher education. The programme will also be suitable Water Reclamation Technologies
for graduates who wish to be involved in the maritime Waste Reuse and Resource Recovery
profession. Environmental Biotechnology
Integrated Urban Water Management
Master of Science (Environmental Science and Environmental Fluid Mechanics
Engineering) Sediment Transport

Coastal Management
The NEWRI Environmental Master of Science (NEMS)
programme is offered by NTU’s School of Civil and Maritime Studies
Environmental Engineering with a Summer attachment Maritime Logistics
Port Economics and Management
at Stanford University, and the Nanyang Environment
Maritime Strategy and Risk Management
and Water Research Institute (NEWRI) on the research
Strategic and Quality Management in Shipping
project component. The programme is designed to prepare
Supply Chain Management
students to be at the forefront of Environmental Engineering
with a combination of coursework and project/research Transportation Engineering
components. It aims to produce high calibre environmental Transport Modelling and Traffic Management
engineers equipped with both fundamental understanding Transport Planning
and practical skills. Congestion Pricing
Road Safety Engineering
10
ACHIEVEMENTS AND COMMENDATIONS
ACHIEVEMENTS AND
COMMENDATIONS
AWARDS (3) Dr Sun was interviewed by NanoGlobe
“Nanostructured Photocatalytic Materials Enable
Professor Chiew Yee Meng was awarded the Distinguished Capturing Solar Energy and Simultaneously Powering
Contribution to Sediment Research in 2010 by World Water Purification” - An interview of Associate
Association for Sedimentation and Erosion Research Professor Darren Delai SUN, Nanyang Technological
(WASER) for the publication of an outstanding research University, Singapore
paper in International Journal of Sediment Research entitled http://www.nanotech-now.com/columns/?article=474.
“Scour Protection around Bridge Piers with Tetrahedron
Frames”. (4) “Water and Industrial 2009 Conference”, Palmerston
North New Zealand, 30 November to 2 December 2009.
Associate Professor Darren Sun’s past research has yielded http://seat.massey.ac.nz/conferences/water09/#.
major achievements in the development of mesoporous Dr Darren Sun was a Co-Chair for this conference
nanostructured Titanium Dioxide microspheres and the
free-standing, flexible and multifunctional Titanium Associate Professor Leong Eng Choon was awarded
Dioxide nanofiber/tube filtration membrane. These new the Singapore Accreditation Council Assessor Award
generation membranes will be able to produce high quality (Distinguished) and SPRING Singapore Merit Award in
drinking water and concurrently generate electricity, thus 2010.
eliminating the cost of water production and chemical usage
together with waste minimization. His innovation has a Assistant Professor Tang Chuyang received the 2010
major impact in membrane technology and many venture Fellowship Award from the International Desalination
capitalists have shown keen interest to commercialize his Association (IDA) for his work on membrane technology
innovation. Associate Professor Sun has 5 patents on for water reuse and desalination.
Titanium Dioxide membrane field. Since 2006, he has
attracted research grants totaling S$17.6 million. He has Assistant Professor Yang Yaowen received the 2009
won several prestigious awards in the past including the Teacher of the Year Award at the Nanyang Awards 2009.
International Water Association Innovation Award 2008,
the IES Prestigious Engineering Achievement Award 2008,
and the Enterprise Challenge Award 2006.
(1) Professor Sun’s Research on TiO2 solar cell was

reported by The Straits Times on 18 April 2009. PATENTS
(2) Dr Sun was interviewed by Sandra Upson, Associate Sun, Darren Delai, Yinjie Wang, Liu Jincheng and Xiwang
Editor, IEE Spectrum on Water and Energy issue related Zhang (2009). “Concurrent electricity and clean water
to research. IEEE Spectrum, June 2010, Page 56. production module made by dye sensitized titanium dioxide
Singapore’s Water Cycle Wizardry: Singapore’s nanostructures”. Ref No TECH/080/09, TD/080/09.
toilet-to-top technology has saved the country from
shortages-and a large electricity bill.
11
ACHIEVEMENTS AND COMMENDATIONS
INVITED LECTURES • Invited Lecture: “Recent development in the

construction of coastal structures” at the Kyoto
Professor Pan Tso-Chien was invited as a speaker for the Seminar 2010 on Geotechnics/Earthquake
following conferences: Geotechnics towards Global Sustainability, January
12-14, 2010, Kyoto, Japan.
• Keynote Paper: “An overview of the current
research programmes in Protective Technology • Invited Lecture: “Strain softening and instability of
Research Centre at NTU.” Proceedings of the 3rd sand and practical application” at the 2010 Huang
International Conference on Design and Analysis Wenxi Lecture, Chinese Institution of Geotechnical
of Protective Structures 2010 (DAPS-2010), 10-12 Engineering, 11 April 2010, Nanjing, China.
May 2010, Singapore, pp. K25-K39. • Invited Lecture: “Improvement of ultra soft soil for
• Keynote Speech: “Developing technology for the reclamation of a slurry pond in Singapore” at
protection.” Inaugural Workshop on Building the Symposium on New Techniques for Design and
Infrastructure Protection for Homeland Security, Construction in Soft Clays, 22 and 23 May 2010,
13 May 2010, Singapore. Brazil.
• Keynote Lecture: “Seismic hazard of low/moderate • Keynote Lecture: “Land reclamation and related soil
seismicity regions – Singapore’s perspective.” The improvement methods in Singapore” at the Australia
10th International ROSE School Seminar, 20-21 May Geomechanics Society Ground Improvement
2010, EUCentre, Collegio Cardinale Riboldi, Pavia, Workshop, 11-12 June 2010, Perth, Australia.
Italy. • Keynote Lecture: “Recent development in ground
improvement methods” at the 7th International
Professor Harianto Rahardjo and Associate Professor Conference on Ground Improvement Techniques,
Leong Eng Choon were invited to deliver a keynote lecture 23-25 June 2010, Seoul, Korea.
entitled “Laboratory characterisation of unsaturated soil for
• Keynote Lecture: “Methods for construction of
slope stability studies” at the 4th Asia-Pacific Conference on
coastal protection structures” at the Conference
Unsaturated Soils, Newcastle, Australia, 23-25 November
on Natural Hazards and Countermeasures in
2009, pp. 565-578.
Geotechnical Engineering, 4-5 November 2010,
Dhaka, Bangladesh.
Associate Professor Darren Sun was invited as a speaker
for the following International Conferences • Keynote Lecture: “Methods for the improvement of
high water content soft clay and sewage sludge”
• Special Invited Speaker: “TiO2 Nanofiber/tube
at the International Symposium, Exhibition, and
Membrane Powering the Water and Energy
Short Course on Geotechnical and Geosynthetics
Productions” at IWA Leading Edge Technology
Engineering: Challenges and Opportunities on
Conference at Singapore International Water Week,
Climate Change, 7-9 December 2010, Bangkok,
June 2009.
Thailand.
• Invited Keynote Speaker: “Smart Multi functional
TiO2 nanotube membrane for Water and Energy
Production” at IWA Water and Industry 2009, New
Zealand, 30 November to 2 December 2009.
Associate Professor Chu Jian was invited to deliver

lectures at the following international conferences:
• Keynote Lecture: “Innovative dike construction
methods” at International Symposium on Geotechnical
Engineering, Ground Improvement & Geosynthetics
for Sustainable Mitigation and Adaptation to Climate
Change including Global Warming, 3-4 December

2009, Bangkok, Thailand.
12
RESEARCH CENTRES
Activities of Centre for Infrastructure Systems (CIS)

from August 2009 to July 2010
(A) Centre Activities Speakers:
• Assistant Professor Jasmine Lam
Seminars, Short Courses & Symposium • Mr Gunasagaran, PSA
• Dr W. Y. Yap, ITMMA
1. BI and NTU Joint Programme –
An Intensive 5-Day Short Course on “Key Elements 3. Public Seminar on “Bearing Capacity of Uplift
of Shipping” Piles Under Deep Excavation”
CIS and BI jointly organized a five-day intensive short A public seminar on “Bearing Capacity of Uplift Piles
course on “Key Elements of Shipping” from 9–13 Under Deep Excavation” was held on 8 December
March 2010. It was attended by 25 participants from 2009. Professor Maosong Huang from the Department
government agencies and the maritime industry, such of Geotechnical Engineering at Tongji University,
as professionals from the Maritime and Port Authority Shanghai, China, was the speaker of the seminar.
of Singapore (MPA), Mediterranean Shipping Company His general research interests are focused on strain
(MSC), I.M. Skaugen and J.B. Ugland Shipping. localization and progressive failure in soils, strength
and deformation characteristics of soft clays, traffic-
Speakers: load-induced permanent deformation, constitutive
modeling of soils, pile-supported earth platform,
• Assistant Professor Jasmine Lam centrifuge modeling, stability analysis of soil structures,
• Cathrine Bjune, BI Norwegian School of and pile foundations in soft clays.
Management
• Captain Robert Gordon, Seasia P&I Services
• Professor Barry Dubner, Barry University
Andreas
• Lasse Rochstad Lim, Tuffchem Shipping Ltd.
• Stephen Fordham, Wikborg Rein
• Mike Pollen, K&L Gates
• Lewis Hart, Willis
• Dag Olav Halle, DNV
4. Invited Public Lecture on “Reinforced Soil System:

Bring Research to Applications”
Associate Professor Robert Lo from the University
of New South Wales, Australia is the invited speaker
of CIS on 17 May 2010 who delivered a public
lecture on “Reinforced Soil System: Bring Research
to Applications”. He was awarded the Thomas Telford
2. Executive Programme on “Port Management and Prize for his paper on the topic of partial factors
Finance Programme” in geotechnical design and was a member of TC-9
CIS organized a two-day short Executive Programme on (reinforced soil) and a core member of TC-39 (coastal
Port Management and Finance from 29–30 July 2010. disaster mitigation). Professor Lo has published over
It was attended by 28 participants from government 140 research articles in soil behaviour, liquefaction,
agencies and the port and maritime industry, such as soft clay engineering, reliability analysis and limit
professionals from the Maritime and Port Authority state design, pavement geotechnics, and reinforced
of Singapore (MPA), Jurong Port and V Ships. soil technology.
13
RESEARCH CENTRES
5. Joint Public Seminar on “Improving the Productivity To develop a new and economical construction material,
and Peformance of Social Infrastructure Project biocement, using the latest microbial biotechnologies; 2)
Delivery” and Public Private Partnership (PPP) To develop cost-effective and environmentally friendly
Success Stories microbiological methods to use biocement for geotechnical
or environmental engineering problems. These include
constructions for roads, tunnels, land reclamation, slope
stabilization, shore protection, and waste treatment; and
3) To study the fundamental principles, microbiological
and biochemical mechanisms that govern the formation
of biocement by microorganisms. So far, suitable
microorganisms and nutrients that could be used for making
different types of biocements were identified. The properties
of soil and waste before and after the treatment using
biocement were studied. Some patents applications have
also been reviewed. Methods for creating water pond in
sand and for the mitigation of liquefaction potential have
been developed.
The afternoon of 15 July 2010 witnessed the Research Project on Planning and Management of
presentation of Infrastructure Projects and Public- Infrastructure Systems Phase I: Studies on Mega
Private-Partnership, by two professors from Curtin Projects in Singapore
University of Technology, Australia and Nanyang
The project is focusing on the planning and management
Technological University, Singapore which was held at
issues related to implementation of mega infrastructures
the Singapore campus of Curtin University, located in
in Singapore. It researches into Lifecycle Management
Jalan Rajah. Professor Love is from the School of Built and consists of coordinating planning, designing, building,
Environment, Art and Design, Faculty of Humanities operation and maintenance of facilities to achieve
at Curtin University in Perth, Western Australia. He management competitiveness. The project objectives
delivered a topic on “Improving the Productivity and are to create a systems-level approach in building and
Performance of Social Infrastructure Project Delivery” managing a single portfolio of systems in Singapore, to
and discussed three social infrastructure projects create an infrastructure database and lay the foundation
in Western Australia, which included a hospital, a for further studies on the lifecycle management of these
school and a prison. The second presentation was mega projects. It uses the system dynamics methodology
by Associate Professor Robert Tiong, School of to propose a strategy of operational & economic success
Civil & Environmental Engineering and Deputy of these mega infrastructures. The project is currently at
Director, Centre for Infrastructure Systems, Nanyang its stage I phase where it is focusing on MRT as the pilot
Technological University, Singapore. Associate infrastructure project to be studied. Initial data collection
Professor Robert Tiong focused on success stories on its planning & design, construction, extension projects
of Public-Private Partnerships (PPP) in this region, and future project has been done. Going forward, the
with a key focus on China, India, Korea, Thailand, challenges in this research would be in data collection and
Taiwan, Singapore, Malaysia and the Philippines. He obtaining more information from government and agencies
also elaborated on typical causes of PPP failures. that are involved in the mega infrastructure projects in
Singapore.
(B) Research and Development Research Project on Study of Transport Energy

Efficiency, Methodology, Practice and Policy Effect
CRP Grant on Underwater Infrastructures and This 2-year research project, which is undertaken by NTU-
Underwater City of the Future CIS in collaboration with the Land Transport Authority,
Centre Director, Associate Professor Chu Jian, together with started in September 2010. The aim is to study the issue
Associate Professor Susanto Teng, Associate Professor Tan of road transport energy efficiency (EE) in Singapore.
Soon Keat and Centre Deputy Director Associate Professor The project investigators are Assistant Professor Chang
Wei-Chung Victor, Associate Professor Gopinath Menon,
Robert Tiong, are awarded a $10M grant by the National
Associate Professor Wong Yiik Diew, Project Officer Ms
Research Foundation (NRF) for the above project.
Lu Ping, PhD Candidates Ms Ho Sze Hwee and Mr Ho
Sijie. The project encompasses literature survey of EE best
practices and estimation models, assembly of databases
Research Projects related to fuel efficiency based on indigenous data from
stake-holder agencies and field surveys, and development of
Research on Biocement – A new sustainable and energy fuel consumption models for road vehicles. The study will
saving material for construction and waste treatment also look into fuel management practices of fleet operators
The main objectives of this inter-discipline study are 1) as well as conducting possible field trials.
14
RESEARCH CENTRES
Research Project on Mobility of Visually-Handicapped 3) 2010 Huang Wenxi Lecture, annual lecture series
Pedestrians – Crossing Behaviour and Assistive Design/ organised by the Chinese Institution of Geotechnical
Technologies at Signalised Pedestrian Crossings Engineering, 11 April 2010, Nanjing, China.
Mitsui Sumitomo Insurance Welfare Foundation (MSIWF) Associate Professor Chu Jian was invited to this event
awarded a research grant to Associate Professor Wong Yiik to deliver an Invited Lecture.
Diew, Associate Professor Gopinath Menon, and doctoral
student Ms Koh Puay Ping to research into traffic safety 4) 1st GlobalTech Workshop on Sustainable Urban
of visually-handicapped pedestrians (VHPs) at signalised Solutions organized by Shanghai Jiao Tong University
pedestrian crossings. The 12-month project is aimed at and the Global Alliance of Technological Universities,
developing a better understanding of VHP’s crossing 3-4 May 2010, Shanghai, China.
behaviour at signalised pedestrian crossing facility, and Associate Professor Chu Jian attended this workshop
studying ‘best practices’ in the provision of assistive as a delegate from NTU and delivered a lecture.
design/technologies. The study shall contribute towards
enhancing the mobility of VHPs on the roads. 5) Symposium on New Techniques for Design and
Construction in Soft Clays, Brazil, 22-23 May 2010.
Associate Professor Chu Jian was invited to this
conference to deliver an Invited Lecture.
International Conference Participation
6) AGS Ground Improvement Workshop organised by
1) International Symposium on Geotechnical Engineering,
the Australia Geomechanics Society, 11-12 June 2010,
Ground Improvement & Geosynthetics for Sustainable
Perth, Australia.
Mitigation and Adaptation to Climate Change including
Associate Professor Chu Jian was invited to this event
Global Warming, 3-4 December 2009, Bangkok,
to deliver a Keynote Lecture.
Thailand.
conference to deliver a Keynote Lecture.
2) Kyoto Seminar 2010 on Geotechnics/Earthquake

Geotechnics towards Global Sustainability, 12-14
January 2010, Kyoto, Japan, organized by Kyoto
Sustainability Initiative (KSI), Kyoto University.
conference to deliver an Invited Lecture.
15
RESEARCH CENTRES
Activities of DHI-NTU Centre in 2010

RESEARCH FOCUS/RESEARCH universities worldwide has been actively pursued, and joint
projects has been initiated together with local governmental
HIGHLIGHTS agencies or industry as described in the following. A
highlighted event was that in September 2010, NEWRI
The DHI-NTU Centre is a joint collaborative effort between
signed a MoU with NPark on promoting research and
NTU and DHI, Denmark (previously known as Danish
development, and together the Centre also signed a project
Hydraulic Institute). The Centre initiated operation in
agreement with NPark on the development of a modular
October 2007 with funding support from the Environment
bio-retention system for urban storm water management.
and Water Industry Development Council (EWI), Singapore
National Research Foundation. It is an integral part of the
In June 2010, Assoc Professor Adrian Wing-Keung Law
ecosystem of the Nanyang Environment and Water Research
was appointed as co-Director of the Centre (taken over
Institute (NEWRI) in NTU. The mission of the Centre is
from Assoc Professor Tan Soon Keat who had assumed
to advance research in the following areas:
the role of Deputy Director, NEWRI), together with Dr
(a) Urban Planning and Water Management Ole Larson, the co-Director from DHI.
(b) Industrial Water Management
(c) Solid Waste Management
(d) Environmental Impact Assessment RESEARCH OUTPUT
(e) Decision Support System Tools and Technologies
1) Research Publications in Year 2010
In Year 2010, the Centre had accelerated the development
of research projects in the above areas. Cooperation with
a) Refereed Journals
Title of paper Author name(s) Journal title, Vol. no. Date
A comparison of municipal solid waste Zhang, D.Q., Tan, S.K., R.M. Waste Management, 30, 921-933 2010
Gersberg, management in Berlin and
Singapore
Urban solid waste management in Zhang D.Q., Tan, S.K., Journal of Environmental 2010
China: status, problems and challenges Gersberg, R.M. Management, Vol. 91, Issue 8,
1623-1633
Extreme Air-gap Response below Deck Li, J., Huang, Z. and Tan, S.K. International Journal of Ocean 2010
of Floating Structures. and Climate Systems, 1(1): 15-26
Lagrangian Modeling of Tidal Bores Jing LI, Huaxing LIU and Journal of Hydrodynamics, 20(5), 2010
Passing Through Bridge Piers Soon Keat Tan supplement: 513-519
Turbulent velocity profiles: a new law Pu J.H., Bonakdari H., La Houille Blanche International 2010
for narrow channels Lassabatere L., Joannis C. and Journal, Vol. 3, pp. 65-70
Larrarte F. (DOI : 10.1051/lhb/2010036)
Google Earth as a tool in 2-D Nguyen Quang Chien and Computers & Geosciences; online 2010
hydrodynamic modeling Tan Soon Keat
A novel application of a Amin Talei, Lloyd H.C. Chua Expert Systems With Applications, 2010
neuro-computational technique in and Chai Quek Vol: 37 (2010) 7456-7468
event-based rainfall-runoff modeling
Evaluation of rainfall and discharge Amin Talei, Lloyd H.C. Chua Journal of Hydrology, Vol: 391 2010
inputs used by Adaptive and Tommy S.W. Wong (2010) 248-262
Network-based Fuzzy Inference Systems
(ANFIS) in rainfall-runoff modelling
Experimental and numerical study on Gao Y.-Y., Etienne S., Fluid Dynamics Research, Vol. 42, 2010
flow behavior behind two unequal Yu D.-Y., Tan S.K., Wang X.K doi:10.1088/0169-5983/42/5/055509
circular cylinders in tandem arrangement and Hao Z.
16
RESEARCH CENTRES
b) Conferences
Title of paper Author name(s) Journal title, Vol. no. Date
Gray water treatment in an Urban Area Zhang D.Q., Tan, S.K. and Proceeding, 4th International 2010
of Beijing, China Gersberg, R.M. Conference on Bioinformatics
and Biomedical Engineering,
Vol. 6 Water Quality &
Public Health – Treatment,
Disposal and Discharge of
Wastewater, Chengdu, China
A case study of silt screen performance Vu T.T., Tan S.K. and Proceeding, 19th World Dredging 2010
Doorn-Groen S. Congress, Beijing, China
Rehabilitating Urban Water System for Dong X., Du P.F. and Proceeding, Conference in Urban 2010
the Inner City of Beijing: Status and Zeng S.Y. Environment Pollution, Boston,
Challenges USA
Modelling of flow in Everglades Nguyen Quang Chien and Proceeding, 17th Asian Pacific 2010
National Park, Florida, USA using a Tan Soon Keat Division Congress, International
quadtree grid Association of Hydro-Environment
Engineering and Research,
Auckland, New Zealand
Near field mixing process of multi-port Nguyen Quang Chien and Proceeding, 6th International 2010
diffusers: numerical modelling with Tan Soon Keat Symposium Environmental
quadtree grids Hydraulics, Athens, Greece
Hydrodynamics of trapezoidal Wang X.K., Hao Z. and Proceeding, 9th International 2010
embankment weirs Tan S.K. Conference on Hydrodynamics,
Shanghai, China
Environmental fluid dynamics – jet flow Wang X.K. and Tan S.K. Proceeding, 9th International 2010
Conference on Hydrodynamics,
Shanghai, China
Laboratory investigation of hydraulic Vu T.T. and Tan S.K. Proceeding, 9th International 2010
performance of silt screen Conference on Hydrodynamics,
Shanghai, China
Three dimensional Simulation of Bore Liu, H. and Tan S.K., Li, J. Proceeding, 29th International 2010
Flow using SPH. Conference on Ocean, Offshore
and Arctic Engineering, Shanghai,
China
Experimental studies of vortex structures Hao Z., Zhou T., Wang X.K. Proceeding, 29th International 2010
in the wake of a cylinder with helical and Tan S.K. Conference on Ocean, Offshore
strakes and Arctic Engineering, Shanghai,
China
Flow around a pipeline near a smooth Wang X.K., Hao Z. and Proceeding, 29th International 2010
bed in steady current Tan S.K. Conference on Ocean, Offshore
China
Flow behaviour behind two side-by-side Gao Y.-Y., Yu D.-Y., Tan S.K., Proceeding, 29th International 2010
circular cylinders with unequal diameters Wang X.K. and Hao Z. Conference on Ocean, Offshore
China
Quantification of viable Enterococcus Goh, S.G. and Gin, K.Y.H. Proceeding, 14th International 2010
faecalis in recreational water by propidium Biotechnology Symposium, Italy
monoazide quantitative PCR
The use of Adaptive Network-based Amin Talei, Lloyd H.C. Chua Proceeding, 9th International 2010
Fuzzy inference System (ANFIS) in and Chai Quek Conference on Hydroinformatics,
event-based rainfall-runoff modeling Tianjin, China
Velocity profiles for shallow, vegetated Nguyen Hoai Thanh Proceeding, International 2010
open channel flows Conference on Fluvial Hydraulics,
Germany
Three-Dimensional Scour at Submarine Wu, Y. and Chiew, Y. M. Proceeding, 5th International 2010
Pipelines in Unidirectional Steady Current Conference on Scour and Erosion,
7-10 November, San Francisco,
USA, Geotechnical Special
Publication No. 210, ASCE.
Effect of Seepage on River Bank Stability Chiew, Y. M., Narasimhan, N. Proceeding, 5th International 2010
and Chu, J. Conference on Scour and Erosion,
7-10 November, San Francisco,
USA, Geotechnical Special
Publication No. 210, ASCE.
Lateral dispersion of granular flows down Cheng Nian Sheng 2010 International Debris Flow 2010
a rough plane Workshop, Chengdu, China
17
RESEARCH CENTRES
COLLABORATIONS AND c) National Research Fund (NRF): on hydraulics

calculation and environmental assessment of
PARTNERSHIPS underwater city structure.
1) Collaboration with universities: d) Temasek Defence Systems Institute (TDSI):
on study of integrating flexible dye sensitized
a) Asia Pacific: Ocean University (China), Tsinghua solar cells into flexible sheets for environmental
University (China), Changsha University of Science applications.
and Technology (China), Sichuan University
(China), University of Peradeniya (Sri Lanka), e) National Parks Board (NParks): on developing
University of Wollongong (Australia), University a modular bio-retention system for urban storm
of Western Australia – to initiate projects on water water management.
quality and watershed management. f) SembCorp Utilities: on studying a virtual brain
b) US: San Diego State University – to conduct system for treatment processes and performance
experiments on constructed wetland for forecasting of an anaerobic biological treatment
pharmaceutical wastewater. system.
c) UK: University of Bradford, University of
Birmingham – to apply modelling on urban storm
water management. EVENTS, CONFERENCES &
SYMPOSIA
2) Projects with local government agencies and
industry: The Centre assisted in the organization of the International
a) Maritime and Port Authority of Singapore (MPA): Conference on Vision and Roadmap for R&D Priorities in
on sediment transport in the presence of silt Maritime Environment, Technology, Business, Policy and
screen. Security by NTU in June 2010.
b) Land Transport Innovation Fund (LTIF): on
assessing environmental risk of CNG vehicle.
18
RESEARCH CENTRES
Activities of Maritime Research Centre (MRC)

from 2009 to 2010
Maritime Research Centre (MRC) has established itself simulations”. The 5th International Conference on
successfully as the bridge between the maritime community Asian and Pacific Coasts, Singapore, 14-16 October.
and the research community in NTU. MRC has strengthened
the link with the Maritime and Port Authority of Singapore 4. Wang, X.K. and Tan, S.K., 2009. “Experimental
(MPA) and Economic Development Board (EDB), IE study of flow about a square cylinder placed on a
Singapore, and has established new and strong working wall”. Proceedings of the 8th International Symposium
relationship with the American Bureau of Shipping (ABS), on Particle Image Velocimetry – PIV09, Melbourne,
Sembcorp, Keppel and other offshore engineering companies Victoria, Australia, 25-28 August.
such as Acergy. The centre is active in outreach activities
and establishes contact with local and regional institutes 5. Wang, X.K., Hao, Z. and Tan, S.K., 2009. “Wavelet
and companies. Analysis of Flow Images Obtained by PIV (Particle
Image Velocimetry)”. Civil Engineering Research,
Vol. 22, pp. 43-46.
New Projects (2009-2010)
6. Kurniawan, A. and Ma, G.W., 2009. “Optimization
of ballast plan in launch jacket load-out”. Structural
1. Sediment transport in the vicinity of Silt-screen
and Multidisciplinary Optimization, Vol. 38, pp. 267-
288.
2. Development of portable on-site sulphur detection
device for bunker
7. Lie, S.T., Yang, Z.M. and Gho, W.M., 2009. “Validation
of BS7910:2005 Failure Assessment Diagram for
3. Development of an innovative gangway for vessel-to-
Cracked Square Hollow Section T-, Y- and K-joints”.
platform operation.
International Journal of Pressure Vessels and Piping,
Vol. 86, No. 5, pp. 291-344.
Academic Staff Exchange Programme 8. Lie, S.T. and Yang, Z.M., 2009. “Fracture Assessment
(2009-2010) of Damaged Square Hollow Section (SHS) K-
joint Using BS7910:2005”. Engineering Fracture
1. Shanghai Maritime University Mechanics, Vol. 76, No. 9, pp. 1303-1319.
2. Sichuan University
3. China Ocean University 9. Lie, S.T. and Yang, Z.M., 2009. “Safety Assessment
4. Shanghai Jiao Tong University Procedure for a Cracked Square Hollow Section
(SHS) Y-joint”. International Journal of Advances in
Structural Engineering, Vol. 12, No. 3, pp. 359-372.
Selected publications (2009-2010)
10. Lie ,S.T. and Yang, Z.M., 2009. “Static Ultimate
1. Kurniawan A., Huang Zhenhua, Li Jing, Liu C., Strength of Cracked Square Hollow Section Y-joint”.
Wang X., Hao Z., Tan S.K. and Edwin N., 2009. “A Civil Engineering Research Bulletin, School of Civil
& Environmental Engineering, Nanyang Technological
numerical analysis of the response and air gap Demand
University, Singapore, pp. 94-96.
for Semi-submersibles”. Proceedings of the 29th
International Conference on Offshore Mechanics and
11. Lie, S.T. and Zhang, B.F., 2010. “Plastic collapse load
Arctic Engineering (OMAE2009), Honolulu, Hawaii,
investigation for safety assessment of cracked square

USA, 31 May-5 June.
hollow section (SHS) T-, Y- and K-joints”. OMAE
2010, Shanghai, OMAE2010-20324.
2. Jing Li, Soon Keat Tan, Zhenhua Huang and Adi
Kurniawan, 2009. “Wave Amplification and Air-gap
12. Low, Y.M. and Grime, A.J., 2010. “Extreme
Response under a Multi-column Platform”. Conference
response analysis of floating structures using coupled
of Coastal Dynamics 2009, Tokyo, Japan, 7-11 frequency domain analysis”. OMAE 2010, Shanghai,
September. OMAE2010-20033.
3. Huang, Z.H., Liu, C.R., Kurniawan, A., Tan, S.K.
13. Li, F.Z. and Low, Y.M., 2010. “Sensitivity study
and Nah, E., 2009. “Responses of a free-floating
of critical parameters influencing the uncertainty of
rectangular caisson to regular waves: Comparisons of
fatigue damage in steel catenary risers”. OMAE 2010,
measurements with time-domain and frequency-domain
Shanghai, OMAE2010-20045.
19
RESEARCH CENTRES
14. He, J.W. and Low, Y.M., 2010. “Probabilistic 23. Jing LI, Huaxing LIU and Soon Keat Tan, 2010,
assessment of the clashing between flexible marine Lagrangian modelling of tidal bores passing through
risers”. OMAE 2010, Shanghai, OMAE2010-20046. bridge piers, Proceedings of the 9th International
Conference on Hydrodynamics (ICHD – 2010), 11-15
15. Wang, X.K., Hao, Z. and Tan, S.K., 2010. “Flow October 2010, Shanghai, China, pp. 513-519.
around a pipeline near a smooth bed in steady current”.
Proceedings of the 29th International Conference on
Ocean, Offshore and Arctic Engineering (OMAE2010), Workshops-seminars-conference (2009-2010)
6-11 June, Shanghai, China, OMAE2010-20749.
1. Public Seminar on “Regional Environmental Simulator
16. Hao, Z., Zhou, T., Wang, X.K. and Tan, S.K., 2010. (RES) and its applications”, 6 February 2009.
“Experimental studies of vortex structures in the wake
of a cylinder with helical strakes”. Proceedings of the 2. Training course: “In-house Training Course for HDB
29th International Conference on Ocean, Offshore and - Land Reclamation and Coastal Protection Work -
Arctic Engineering (OMAE2010), 6-11 June, Shanghai, Design and Analysis”, February-March 2009.
China, OMAE2010-20181.
3. Public Seminar on “Important role of R&D in offshore
17. Gao, Y.-Y., Yu, D.-Y., Tan, S.K., Wang, X.K. and EPCI contract”, 27 May 2009, NTU
Hao, Z., 2010. “Flow behaviour behind two side-
by-side circular cylinders with unequal diameters”. 4. Public Seminar on “Panama Canal Third-lane Locks
Proceedings of the 29th International Conference on and Access Channel Expansion Program”, 8 June 2009,
Ocean, Offshore and Arctic Engineering (OMAE2010), NTU
5. 5th International Conference on Asian and Pacific
18. Huaxing Liu, Soon Keat Tan, Jing Li and Xikun Wang, Coasts (APAC2009), 13-16 October 2009, NTU
2010. “Three dimensional simulation of bore flow (Singapore)
using SPH”. Proceedings of the 29th International
Conference on Offshore Mechanics and Arctic 6. NTU – SJTU International Workshop on R & D in
Engineering (OMAE2010), Shanghai, OMAE 2010- Civil and Environmental Engineering. 12 October 2009
21090. (Singapore)
19. Chunrong Liu, Zhenhua Huang, Adrian Law Wing 7. Training Course for BCA - An Introductory Course
Keung and Nan Geng, 2010. “A Numerical Study of on Coastal Engineering, Analysis-Design-Application,
Wave Energy Converter in the Form of an Oscillating July-August, 2009 (Singapore).
Water Column Device Based on a Mixed Eulerian-
Lagrangian Formation”. Proceedings of the 29th 8. Vision and Roadmap for R&D Priorities in Maritime
International Conference on Ocean, Offshore and Environment, Technology, Business, Policy and
Arctic Engineering (OMAE2010), 6-11 June, Shanghai, Security, 29-30 April, 2010 Hilton Singapore,
China, OMAE2010-21056. Singapore.
20. Nguyen Quang Chien and Tan Soon Keat, 2010. “Near 9. NTU-SMU International Workshop on Offshore
field mixing process of multi-port diffusers: numerical Engineering, June 2010, Shanghai China.
modelling with quadtree grids”. Proceedings of the
International Symposium on Environmental Hydraulics, 10. In-house Training Course for Marine Contractor, July
Athens (in press) 2010, Singapore.
21. Wang, X.K., Hao, Z. and Tan, S.K., 2010. 11. Professional Training Course for HDB on the
“Hydrodynamics of trapezoidal embankment weirs”. Design of Container Bund, October-November, 2010

Proceedings of the 9th International Conference on (Singapore).
Hydrodynamics (ICHD – 2010), 11-15 October 2010,
Shanghai, China, pp. 386-390. 12. Training Course for LTA - Interpretation of Geotechnical
Design Parameters to Geotechnical Design Parameters
22. Wang, X.K. and Tan, S.K., 2010. “Environmental and Laboratory Testing, November 2009, Singapore.
fluid dynamics – jet flow”. Proceedings of the 9th
International Conference on Hydrodynamics (ICHD
– 2010), 11-15 October 2010, Shanghai, China, pp.
1009-1014.
20
RESEARCH CENTRES
Activities of Protective Technology Research Centre

(PTRC) from November 2009 to September 2010
OUTREACH PROGRAMMES moderated strain rates before and beyond damage with
non-linear equation of state properties for strong shock
The outreach programmes provide a platform for knowledge waves. The model is readily available to all users of the
transfer and they also help PTRC to establish collaborations commercial hydrocode AUTODYN and continuously
with local and foreign agencies in the area of protective supported since the year 2000. Over the last decade it
technology and homeland security. There are 5 public has found numerous worldwide applications reflected in
seminars, 1 international conference and 1 workshop publications. They deal with dynamic load cases such
organized by PTRC which are summarized as follows:- as projectile and shaped charge penetration, contact
detonation, internal and external blast loading.
Public Seminars
3. Title: Large Scale Blast Simulator Based
1. Title: Liquefaction of Sand in Plane-Strain, Explosive Gas Loading Methods for
18 December 2009 Structures and Recent Advances in
Speaker: Dr. Dariusz Wanatowski, MSc, PhD, HIT BRPE Lab, 28 April 2010
MSEAGS, MASCE, Nottingham Centre Speaker: Professor Zhang Chunwei, Associate
for Geomechanics, University of Professor, School of Civil Engineering,
Nottingham, United Kingdom Harbin Institute of Technology; Visiting
PI (Initiator): Associate Professor Chu Jian Research Fellow, School of Civil and
Liquefaction of granular soils is one of the most Resource Engineering, The University
rapid and thus one of the most dangerous initiation of Western Australia
mechanisms of landslides. Therefore, it requires the PI (Initiator): Associate Professor Ma Guowei
special attention from civil engineers and researchers. Professor Zhang Chunwei met Associate Professor
The majority of experimental studies on static Ma Guowei’s UTRE group to share about impact
liquefaction and instability of sand have been carried and blast tests, especially on the large scale blast
out under axisymmetric conditions. However, most simulator based explosive gas loading methods for
geotechnical structures such as slopes, embankments, structures, as well as the recent advances in Harbin
and retaining walls are not axisymmetric problems and Institute of Technology Blast Resistance and Protective
can only be simplified into plane-strain conditions. A Engineering Laboratory.
comprehensive experimental study on the liquefaction
of a granular soil under plane-strain conditions is 4. Title: Tunnel Fires and Related Concrete
presented in this talk. Undrained tests on very loose
Technological Issues, 24 June 2010
sand under both plane-strain and axisymmetric
Speaker: Dr. – Ing. Frank Dehn, Executive
conditions were conducted and compared. Instability
Director, MFPA Leipzig GmbH, Leipzig
behaviour of very loose sand under drained conditions
is also discussed. Based on the testing data, a unique Institute for Materials Research and
relationship between the stress ratio of the instability Testing
line and the state parameter is established to enable the PI (Initiator): Associate Professor Tan Kang Hai
triaxial results to be used for plane-strain conditions. A number of devastating fires in tunnels has put the
The experimental data presented in this talk shed new subject of concrete spalling back to many research
light on static liquefaction and instability phenomena of agendas. The sensitivity of concrete towards spalling
granular soils, and should be of interest to academics, has shown to be responsible for severe damages of
researchers, and engineers, or anyone else who may concrete linings inside rock-drilled tunnels and/or
be interested in experimental soil mechanics. shield driven soft-soil tunnels. Many trial tests - mostly
mechanically unloaded and on lab-scale - have been
2. Title: A Review of Concrete Modeling and conducted in order to discover the mechanisms that
Hydrocode Applications, 11 February are responsible for this specific “thermo-physical-
2010 mechanical-chemical phenomenon”. Up till now, the
Speaker: Dr Werner Riedel, Deputy Head of mechanisms of spalling are still rather unknown.
Department “Safety Tech. and Protective Currently used measures, such as an addition of
Structures” at the German Fraunhofer synthetic fibres, turned out to be most effective in
Society, Ernst-Mach-Institute
reducing the spalling and heat ingress to a tolerable
PI (Initiator): Professor Fan Sau Cheong
limit so that the structural integrity of concrete tunnel
The RHT concrete model was developed at Ernst- construction can be still guaranteed. However, to verify
Mach-Institut (EMI) in German 10 years ago. It a more realistic material and structural behaviour tests
combines detailed tri-axial strength descriptions at in full scale - which are (simultaneously) mechanically
21
RESEARCH CENTRES
and thermally loaded - are an appropriate measure. UTRE seminars are one of the regular research
Such an approach has been successfully used for activities of the programme. Research progress of the
several tunnel projects all over Europe (Sweden, The UTRE programme is reviewed on a half-yearly basis
Netherlands, Belgium, Germany and Spain) and in during the seminars. During this seminar, discussions
Oceania (Australia). and information on the development of different
projects were exchanged.
The speaker shared with the participants the gained
experiences out of these tunnel projects, gave some Project Speaker Topic
remarks concerning concrete mix design, mode of A1 Mr Fan Lifeng Rock dynamic testing
testing, interpretation of test results and concluded B3 Ms An Xinmei Development on 2D
with recommendations for the construction of - more numerical manifold method
or less - fire proof concrete tunnels. (NMM)
A2 Mr Muley Pravin Continuous structural health
Sudhakar monitoring system-BOTDR
B1 Mr Bao Huirong Implementation of coupled
FEM and DDA approach in
rock modelling
C Associate Professor The Digital Rock
Tor Yam Khoon Engineering System based
on 3DGIS Technology
D Associate Professor Survival condition in
Tan Kang Hai underground fire
Conference
Seminar on Tunnel Fires and Related Concrete Technological The Third International Conference on Design and
Issues (Associate Professor Tan Kang Hai presenting a token Analysis of Protective Structures in 2010 (DAPS2010),
of appreciation to Dr.-Ing. Frank Dehn.) 10 – 12 May 2010
PTRC and DSTA have jointly organized the Third
5. Title: Underground Technology & Rock International Conference on Design and Analysis of
Engineering (UTRE) Phase II Protective Structures (DAPS) 2010. The DAPS Conference
Programme, 9 July 2010 has been well received with a total attendance of around
PI (Initiator): Associate Professor Zhao Zhiye 280 delegates with 34 foreign delegates representing 18
various countries (as compared to 150 in DAPS 2003 and
The UTRE is a joint R&D programme between DSTA
250 in DAPS 2006). In total, 43 technical papers were
and PTRC. Phase II of UTRE is a continuation of
presented during the 3-day conference at Novotel Clarke
the first five-year research efforts contributing to 4
Quay Singapore. Mr Quek Tong Boon, Chief Defence
research areas related to underground technology and
Scientist and Chief Research & Technology Officer from
rock engineering. Phase II has a duration of 4 years
the Ministry of Defence, Singapore delivered the opening
from 2009 to 2013. There are 8 research projects with
keynote speech as the Guest of Honour, while Mr Soh Koh
9 faculty members focusing in different areas.
Pheng (CE DSTA) hosted the conference dinner.
A1: Protection of Underground Structure
A2: Development of a continuous and distributed The conference was jointly organised by Associate Professor
monitoring system for underground applications Tan Kang Hai from PTRC side, and Mr Chua Hian Koon
A3: Risk assessment and management system for from DSTA side.
underground rock cavern projects

B1: DDA modelling and advanced analysis and design The other 4 keynote addresses were presented by Professor
system Ted Krauthammer (University of Florida, USA), Professor
Magnus Langseth (Norwegian University of Science and
B2: Probabilistic stability assessment of rock Technology, Norway), Professor Pan Tso-Chien (Dean,
caverns College of Engineering & Director, PTRC, NTU) and
B3: 2D/3D manifold method modelling for jointed Associate Professor Tan Kang Hai (Deputy Director,
rock mass PTRC, NTU).
C: GIS-based digital rock engineering
D: Tunnel fire safety assessment and evacuation
tool
22
RESEARCH CENTRES
Visit by Dr Lim Chee Onn, NTU Board of

Trustees, Chairman of Singbridge International
Singapore Pte Ltd, 24 May 2010
Third International Conference on Design and Analysis of

Protective Structures (DAPS) 2010
(Mr Quek Tong Boon, Chief Defence Scientist and Chief
Research & Technology Officer from the Ministry of Defence, Visit by Dr Lim Chee Onn
Singapore delivering the Opening Keynote Speech
as the Guest of Honour.)
Visit by Delegates from Ministry of Home
Workshop Affairs, 29 June 2010
Building Infrastructure Protection for Homeland * The delegation was led by Mr Eric Yap, Senior Director,
Security, 13 May 2010 Homefront Security Division, MHA.
In conjunction with the DAPS2010 Conference, DSTA
and NTU organized a half-day workshop “Building
Infrastructure Protection for Homeland Security” on 13
May 2010 at Novotel Clarke Quay. The workshop brought
together interest groups from the respective Ministries
and government agencies to share experiences and create
better awareness on how Singapore may move forward in
developing national capability in protective technology.
Around 75 delegates attended the workshop.
INTERNATIONAL AND LOCAL

VISITORS
PTRC has received 5 delegations (around 70 visitors) during Visit by Delegates from MHA
the reporting period. The visitors whom we have received
include the following:-
Visit by Delegates from the MINDEF, 26 April

2010
* The delegation was led by BG(NS) Ravinder Singh,

DS(T), MHQ.
Visit by Delegates from MINDEF
23
RESEARCH CENTRES
On-going Projects
The table below shows the current projects.
External Collaborating
No Project Title Principal Investigator (s)
Funds (S$) Partners
1 Development of Analytical A/Prof Tan Kang Hai (CEE) 1,260,000 Defence Science and
Tools for Progressive Collapse A/Prof Li Bing (CEE) Technology Agency
due to Terrorist Bombing A/Prof Lee Chi King (CEE) (DSTA)
2 Dynamic Properties of A/Prof Leong Eng Choon (CEE) 164,800 DSTA
Singapore Soils
3 Underground Technology and A/Prof Ma Guowei (CEE) 3,850,000 DSTA
Rock Engineering (UTRE) A/Prof Zhao Zhiye (CEE)
Programme, Phase II A/Prof Yang Yaowen (CEE)
A/Prof Tor Yam Khoon (CEE)
A/Prof Tan Kang Hai (CEE)
A/Prof Chu Jian (CEE)
A/Prof Goh Teck Chee, Anthony
(CEE)
A/Prof Tiong Lee Kong, Robert
(CEE)
Asst/P Wong Ngai Yuen, Louis
(CEE)
4 An Integrated Multiple-Hazards Programme Coordinator: DSTA
Research Programme for Prof Pan Tso-Chien (CEE)
Resilient Structures
Project 1 PIs: 732,930
Effects of Catenary and A/Prof Tan Kang Hai (CEE)
Membrane Actions on the A/Prof Lee Chi King (CEE)
Collapse Mechanisms of RC
Buildings – Behaviours of
Structural Elements
Project 2 PI: 523,600
The Influence of Floor Slabs A/Prof Li Bing (CEE)
and Transverse Beams on the
Behaviour of RC Beam-Column
Joints under Loss of Column
Scenarios
5 Project on Underground Target A/Prof Leong Eng Choon (CEE) 271,200 DSTA
Detection using Ground Tremor
Analysis
6 Prediction of Explosion Prof Fan Sau Cheong (CEE) + 1,564,000 DSTA
Hazards from Earth Covered Collaborators from other Schools in
Magazines College of Engineering
24
RESEARCH CENTRES
Residues and Resource Reclamation Centre (R3C)

The Residues and Resource Reclamation Centre (R3C) Contaminated Site Remediation
was established on 1 May 2009 as a platform for waste
management research, especially on resource recovery and Developing solutions and technologies for remediating
remediation under the Nanyang Environment and Water contaminated sites (e.g. landfill mining, contaminants
Research Institute (NEWRI) of the Nanyang Technological removal from soils and ground water, bio-remediation and
University (NTU) in Singapore. The main objective is utilisation, etc.)
on conducting cutting edge research and strengthening
Singapore environmental industry‘s capability in the area
of waste resource management.
RESEARCH AREAS
Waste to Materials
Converting wastes into new and useful materials (e.g. plastic

waste into biodegradable PHA polymer, incineration bottom
ash into carbonated ash product for high value usage,
nutrient recovery, high quality compost production, etc.)
MISSIONS
R3 Research and Translation in Singapore
and the region
• Acting as a think tank to identify R3 research needs

• Cutting edge interdisciplinary research within NEWRI
and NTU
• Benchmarking with international leading research
institutes
R3 Resource and Technology Transfer

Centre
Waste to Energy
• Research partner of the R3 industry and the public
Harnessing energy from urban biomass, sewage sludge, sector
agricultural residues, micro-algae (e.g. food waste converted • Practical application of research outcome
to hydrogen and methane, bio-ethanol produced from
agricultural and horticultural waste, microbial full cell, • Monitoring and upgrading existing R3 plants
etc.)
Education and Training for R3 Professionals

• Providing research scholarships for PhD programme

• Inviting and working with renowned visiting professors
and scientists
• Organising international workshops, symposia and
conferences
25
RESEARCH CENTRES
CENTRE ACTIVITIES • Site selection for sanitary landfill using GIS

Professor Anil Dikshit
Seminars Centre for Environmental Science & Engineering
(CESE) Indian Institute of Technology Bombay, Powai
• Rehabilitation and remediation of NeiHu landfill Mumbai 400076 India
in Taipei city 25 Jun 2010
Professor Cheng Sheng-Shung
National Cheng Kung University (NCKU) Taiwan • Minimizing environmental impacts of a petroleum
24 Jun 2009 refinery: LCA approach
• Issues, challenges, and opportunities for municipal Centre for Environmental Science & Engineering
solid waste management in Shanghai; (CESE) Indian Institute of Technology Bombay, Powai
Application of aged refuse from landfill mining for Mumbai 400076 India
leachate and feedlot wastewater treatment; 2 July 2010
Biogas production from food waste and cassava
stillage • Recycling science
Assoc Professor NIU Dongjie Dr Peter Rem
UNEP-TONGJI Institute of Environment for Separation Technology Materials & Environment Delft
Sustainable Development (UNEP-IESD) University of Technology The Netherlands
Tonggji University, Shanghai, China 13 July 2010
8 and 15 Jan 2010
• Planning, design and implementation of a solid
• Greenhouse gas and carbon quantification of algal waste managment plan for an urban area
biodiesel Professor Anil Dikshit
Dr Tom Beer Centre for Environmental Science & Engineering
Transport Technologies and Sustainable Fuels Energy (CESE) Indian Institute of Technology
Transformed Flagship, CSIRO Australia Bombay, Powai Mumbai 400076 India
1 Apr 2010 2 July 2010
• Production of biofuels and biochemicals from

renewable resources Colloquium
Dr Wu Jinchuan
Institute of Chemical and Engineering Sciences • Joint colloquium with National Cheng Kung
Singapore University (NCKU) of Taiwan
9 Apr 2010 16 August 2010
A delegate of 6 members consisting of professors
• Sustainable and CO2-free bioenergy production and researchers visited R3C and shared their research
system using lignocellulosic and microalgal interests and findings with R3C researchers.
feedstock
Dr Jo-Shu Chang
Centre for Bioscience and Biotechnology,
Department of Chemical Engineering,
National Cheng Kung University, Taiwan
26 May 2010
• Use of molecular tools for quantifying slowly

growing microbial biomass in environmental bio-
processes
Dr Seokhwan Hwang
School of Environmental Science & Engineering,
Pohang University of Science and Technology
(POSTECH) South Korea
16 Jun 2010
• Arsenic decontamination of groundwater

Centre for Environmental Science & Engineering
( C E S E ) I n d i a n I n s t i t u t e o f Te c h n o l o g y
B o m b a y, P o w a i M u m b a i 4 0 0 0 7 6 I n d i a
18 Jun 2010
26
RESEARCH CENTRES
Workshop News Updates
• Joint ETO/R3C workshop on urban waste • R3C official launch

management 5 October 2009
14 September 2010
It was jointly organized by R3C and Environment • R3C awarded NRF’s competitive research
Technology Office, National Environment Agency programme (CRP) Grant
(ETO, NEA). Eighty participants from government 28 January 2010
agencies, industries and research institutes attended
this workshop. This workshop successfully provided • R3C awarded Environment Technology and
a platform for industries, government agencies, and Research Programme (ETRP) Grant
research institutes to exchange ideas and eventually 8 February 2010
delineated a waste management research road map
for Singapore. • For more information, please refer to:
http://www.ntu.edu.sg/r3c
27
ENVIRONMENTAL AND WATER RESOURCES ENGINEERING
A LABORATORY STUDY OF
WAVE-INDUCED SETUP OVER CORAL
REEFS WITH AN IDEALIZED RIDGE
Yu Yao (yaoyu@pmail.ntu.edu.sg)
Zhenhua Huang (zhhuang@ntu.edu.sg)
Edmond Lo Yat-Man (cymlo@ntu.edu.sg)
S.G. Monismith (monismith@stanford.edu)
ABSTRACT: We report a laboratory study of wave-induced setup over an idealized coral reef that includes a ridge at the seaward
edge -- a geometrical feature commonly found in nature. To understand the role of the ridge in wave-transformation mechanism and
wave-induced setup over the reef, laboratory experiments on a range of water depths and wave conditions were carried out. The focus
of this study is a comparison of wave-induced set-ups obtained with and without the idealized rectangular ridges.
INTRODUCTION flume to generate the designed waves. At the other end, a

beach with a mild slope was covered with porous material
The wave-induced setup due to wave breaking is one to reduce wave reflections. To construct an idealized 2D
of the important factors to consider in determining both fringing reef model, a plane slope of approximate 1:6 was
water level and mass transport above a reef-top, which built with PVC plates at the mid-section 16.35m from the
has ecological as well as engineering significance. It has wave-maker and joined with a horizontal platform which
been observed that a ridge (reef crest) may be present at was 0.35m above the flume bottom. The horizontal platform
the reef edge (Gourlay 1996b, Hench et al. 2008). When was 7m long with its width matching the inner width of
a ridge exists, the strong nonlinear wave-ridge interactions the flume. A rectangular box 55cm long, 50cm wide and
make the problem much more complicated for analytical 5cm high was placed at the edge of the platform to model
analysis. Therefore, a series of experiments were carried a ridge. Over the reef profile, a total of twelve wave gages
were used to measure the water surface elevations. The
out in a wave flume with idealized reef-ridge models being
detailed arrangement is shown in Figure 1. The wave
installed at the reef edge to simulate fringing reefs with
gages have an accuracy of 0.1mm. All the wave gauges
rectangle ridges. Based on our experiments, we will discuss
were sampled at 50Hz by a personal computer through a
wave breaking mechanics, the variation of wave-induced
data acquisition system.
setup, etc.. The comparison between the results with and
without ridge models is made to highlight the effects of The design regular and irregular incident wave conditions
the ridge on the wave dynamics over coral reefs. were selected from a combination of wave heights, wave
periods and water depths (see Table 1 for details). The
irregular waves were generated according to JONSWAP
INSTRUMENT AND EXPERIMENTAL spectra with peak enhancement factor γ = 3.3. During
SETTINGS the experiments, each wave condition was repeated three
times to check the repeatability. After the wave-maker
The laboratory experiments were conducted in a glass- was started, a steady wave condition was assured before
walled wave flume with 36m long, 0.55m wide and starting data acquisition. Between two subsequent tests,
0.60m deep, which is located in the Hydraulic Modeling several minutes were allowed to elapse so that the water
Laboratory at Nanyang Technological University, Singapore. surface could become calm and the effects of residual
A piston type wave-maker was placed at one end of the currents were minimal.
Table 1. Test wave conditions

Range
Type
Without ridge With ridge
Water level (m) 0.35,0.4,0.45 0.4,0.42,0.45
Offshore regular wave height: (m) 0.046-0.131 0.035-0.135
Offshore regular wave period: (s) 0.83-1.67 0.83-1.67
Offshore significant wave height: (m) 0.033-0.091 0.032-0.087
Offshore peak wave period (s): 1.00-1.67 1.00-1.67
28
Figure 1. Sketch of the experimental arrangement
RESULTS frequency portion of wave spectra reflects more efficiently.

The effects of the ridge on the wave reflection are more
Wave Transformation over Reef Crest and Wave evident for waves of small wave slope. The enhanced wave
Reflection reflection is expected since the ridge structure functions like
a submerged breakwater, which is widely used to reflect
Videos of the wave breaking area were taken after the wave the wave energy (Yao et al. 2009).
field had reached a steady state. Plunging breakers were
observed in most of experiments; spilling breakers were
also observed in some of the experiments with water depth
h=0.45m. The breaking point moves from the fore-reef
slope onto the reef-top as water depth increases. Figure 2
shows the representative features of wave transformation
and breaking over the reef crest in the presence /absence
of the ridge for incident regular waves of deep water wave
height Ho=0.095m and wave period T=1.25s in water of
depth h=0.45m. Four different phases are shown in Figure
2, starting from the moment when the lip of breaker hit the
water surface (t/T=0). For the reef without ridge, waves
plunged on the reef flat at t/T=0, while at t/T=1/4, the
splash-up jet due to plunging breaker hit the water ahead
of it, producing an air-water mixture of foam, bubbles and
some subsequent white-capping. After one half wave period Figure 2. Snapshots of regular wave transformation over reef
crest at difference phases (Ho=0.095m,T=1.25s,h=0.45m).
(t/T=1/2), the broken waves propagated across the surf-zone
in the form of fully turbulent bores with a turbulent roller in
the front. For the last phase (t/T=3/4), the bore was mostly Mean Water Level (MWL) across the Reef platform
dissipated and a transmitted wave was reformed on the reef
flat. A strong reverse flow could be observed during this Twelve wave gages (G1 to G12) were used in our
period before the next incoming wave arrived. experiments, enabling us to construct reasonably detailed
setup/set-down profiles across the reef models by a linear
When a ridge was present at the reef crest, the breaking interpolation. Two cases representing regular and irregular
point shifted seaward, the breaking waves strike the front wave conditions are illustrated in Figure 3. As we may
side or the edge of the ridge, and then plunged onto the expect, for both types of waves, set-down occurred for
ridge-top, resulting in stronger wave reflections and air gages located at the seaward side of the surf-zone. Within
entrainment. However, the whole transformation process the surf-zone, there is a monotonic increase of MWL (setup)
is otherwise identical to the reef without ridge. due to the wave attenuation through wave breaking. The
maximum setup always appears at G9; after that point, the

Reflection coefficients for regular waves, which were setup first decreases slightly and then become more or less
determined using a two-probe method (Goda 2000), ranged constant. Under the same wave condition, the magnitude of
from 2% to 25% in the absence of the ridge and from 6% the setup measured in the presence of ridge is significantly
to 55% in the presence of the ridge. For irregular waves, larger than that seen in the absence of the ridge. Moreover,
reflection coefficients ranged from 9% to 37% in the absence the ridge also caused the lowest point of MWL and the
of the ridge and from 20% to 66 % in the presence of point where the waves broke to shift seaward.
the ridge. Irregular waves have larger values of reflection
coefficients than comparable regular waves because the low
29
Under similar wave conditions, the ridge causes a noticeable

increase in the wave setup on the reef-top, particularly
for the cases of longer wave periods. In some cases, the
wave-induced setup is as much as doubled by the ridge.
The wave-induced setup on the reef flat increases almost
linearly with increasing Ho for all cases with and without
a ridge. The relationship between and T is less obvious.
However, by comparing the setup with similar Ho, it is found
that in general increases with increasing T, which agrees
with the observations of Gourlay (1996a). Meanwhile,
the setups for regular waves (Figures 4(a) and 4(b)) are
remarkably larger than those for irregular waves (Figures
4(c) and 4(d)) if we use the offshore significant wave
height (Hso) instead of Ho and offshore peak wave period
(Tp) instead of T to characterize spectral waves.
Figure 3. Mean water level (MWL) offshore and across the

reef profile under different wave conditions: (a) regular waves; CONCLUSIONS
(b) irregular waves. Open circles denote the locations of wave
gages, dash lines indicate reef profile without ridge, and
A series of experiments have been carried out to study the
solid lines indicate reef profile with ridge.
effects of idealized ridge on the wave-induced setup over
a two dimensional horizontal, impermeable reef-top. It is
Wave Setup as a Function of Offshore Wave Height found that the behaviors of the wave transformation in the
presence of a ridge are significantly different from those
The above analysis shows that the MWL always reaches in the absence of the ridge. In particular, the location of
maximum around G9 on the reef platform; thus, the setup at the breaker point can be moved toward deep water side
G9 is a good measure of the maximum setup ( ). Figure 4 and the reflection coefficients of the reef model were
shows this wave-induced setup as a function of deep water dramatically increased. Furthermore, the ridge near the
wave height (Ho), for different wave periods (T) and two reef edge causes a considerable increase in the setup over
water depths of h=0.40m and 0.45m. The cases with and the reef platform under both regular and irregular wave
without a ridge are also compared in the figure. conditions. Finally, it is found that the wave-induced setup
increases with increasing deep-water wave height and wave
period. Analytical and numerical studies will be carried out
to further analyze our experimental data.
REFERENCES
[1] Goda, Y., 2000. “Techniques of Irregular Wave Analysis in

Random Seas and Design of Maritime Structures”. World Sci.
Press, Singapore.
[2] Gourlay, M.R., 1996a. “Wave set-up on coral reefs. 1.
Set-up and wave –generated flow on an idealized two
dimensional reef”. Journal of Coastal Engineering,
27:161-193.
[3] Gourlay, M.R., 1996b. “Wave set-up on coral reefs. 2.
Wave set-up on reefs with various profiles”. Journal of

Coastal Engineering, 28: 17-55.
Figure 4. Maximum wave setup on reef flat as a function of [4] Hench, J.L., Leichter, J.J. and Monismith, S.G.,
deep water wave height for different wave periods, still water 2008. “Episodic circulation and exchange in a wave-
depths, and incident wave conditions. Open markers indicate driven coral reef and lagoon system”. Limnology and
reef profile without ridge, and solid markers indicate Oceanography, 53(6): 2681-2694.
reef profile with ridge.
[5] Yao, Y., Lo, E.Y.M., Huang, Z.H. and Monismith, S.G.,
2009. “An Experimental Study of Wave-induced Set-up
over a Horizontal Reef with an Idealized Ridge” paper
presented at 28th International Conference on Offshore
Mechanics and Artic Engineering, ASME, Hawaii,
USA.
30
ADOPTION AND ACCEPTANCE OF

CNG VEHICLES ON THE URBAN
ENVIRONMENT
Isaac Sadikin (isadikin@ntu.edu.sg)
Lie Seng Tjhen (cstlie@ntu.edu.sg)
ABSTRACT: This study addressed the issue of society perception of CNG as an alternative fuel in a motor-vehicle. The impact of the
adoption of CNG includes environmental impact with respects to fueling station (risk of leakage, stock-pile, increased traffic emission
and noise), incidents of leakages, sudden explosion. The deliverables of the study includes the appropriate measure to address citizens’
concern of close vicinity to stock-pile of CNG in the residential neighborhood, sudden release of CNG to the environment and risk, and
suggested appropriate counter measures.
INTRODUCTION in minimizing pollution from the transportation sector

intensifies.
CNG is an alternative fuel for motor-vehicles besides
using conventional fuel such as petrol, diesel and LPG. Table 1. Adoption of CNG in world wide as of 2008
This alternative fuel had been implemented worldwide CNG (as of 2008)
for the cars, truck, school buses, and trains. The study on Country Number of Vehicles Refueling
the environmental impact, incidents and sudden explosion Stations
of CNG cylinder and refueling station will be discussed Argentina & > 3 million 3352
in this article. Brazil
US 110,000 1,100
Colombia 257,468 378
CNG PROPERTIES & EMISSIONS
Egypt 63,000 95
CNG or compressed natural gas is a domestically available, Iran 1.3 million 750
economical, clean burning, alternative fuel source for Italy The 4th country in the world 800
vehicles. Rather than burn gasoline or diesel fuel, a for number of CNG-powered
vehicles
consumer would fuel their vehicle with natural gas. In
order to provide enough range, the gas is compressed and Sweden 14,500 90
stored on the vehicle in pressurized tanks. CNG tanks can Singapore 5000 5
hold up to 3,600 psi. The fact is that natural gas is a much
safer fuel than gasoline.
INCIDENT
CNG also produces significantly lesser emissions of
pollutants such as carbon dioxide (CO2), hydrocarbons There are a few incidents occurred in several countries
(UHC), carbon monoxide (CO), nitrogen oxides (NOx), due to cylinder ruptures, tank fires or fire exposure, illegal
sulfur oxides (SOx) and particulate matter (PM), as compared use of the cylinder. Most of the incidents did not result in
to petrol. For example, an engine running on petrol for 100 human casualties.
km emits 22,000 grams of CO2, while covering the same

distance on CNG emits only 16,275 grams of CO2. The In Singapore, there is only one incident to date which
corresponding figures are 78 and 25.8 grams respectively, occurred on a bus in August 2010. The investigation of
for nitrogen oxides. Carbon monoxide emissions are the incident is ongoing (see Figure 1 for a media report
reduced even further. Due to lower carbon dioxide and of the incident).
nitrogen oxides emissions, switching to CNG can help
mitigate greenhouse gas emissions. The ability of CNG to
reduce greenhouse gas emissions over the entire fuel life SAFETY & RISK
cycles will depend on the source of the natural gas and
the fuel it is replacing. Consequence analysis was done by considering accident
locations and computing the physiological damage and
The number of CNG vehicles and refueling stations lethality effects of heat fluxes generated from fires. The
are growing from year to year even as ongoing efforts
31
total risk was determined by summing the risk associated

with each fire/accident scenario.
The projected fatality resulting from an unconstrained fire

is 2.2E-5/bus/year. For the 8500 CNG buses in operation
in year 2001 in the United States, this would translate to
approximately 0.19 deaths/year or a mean time to occurrence
of a fatality of 5.4 years/fatality.
If all of the present school buses in the United States are

converted to CNG type, then the projected mean fatality
would be 9.9/year or a mean time to occurrence of a The Straits Times, Saturday, 14 August 2010
fire related fatality of 1.2 months/fatality. Accordingly,
Figure 1. First CNG Accident in Singapore.
catastrophic bus-related failure event leading to a fire is
certainly a major safety issue in CNG powered buses. The
table below summarizes major results of this study. used to describe how the material was dispersed to some
concentration levels. Then, fire and explosion models
Comparing the estimated results for CNG buses with those converted the source model information on the release into
of historical diesel school experience, one may conclude hazard potentials such as thermal radiation and explosion
that CNG buses are on the average 2.5 times more prone to overpressures [2, 3, 4]. All of the mentioned steps were
fire fatality risk than diesel buses. While these comparative modeled using PHAST 6.5 software package developed by
values are based on best estimate averages, the worst DNV. Finally, effect models converted results obtained by
case fire scenarios for CNG buses are expected to lead to software into effects on people represented by probability
higher fatalities as compared to worst case fire scenarios of death. Probit equations are commonly used to quantify
of diesel buses. the expected rate of fatalities for the exposed population.
Finally, risks of non-fatal fire scenarios (those primarily All of the selected scenarios had been investigated in two
leading to injuries) should also be estimated. different atmospheric conditions corresponding to day and
night as detailed in Table 2.
CASE STUDY IN IRAN – RISK ASSESSMENT Table 2. Atmospheric conditions responding to day and night
OF SITING THE CNG REFUELLING STATION Parameters Day Night
(Sharif University of Technology) Wind velocity (m/s) 2.5 2.1
Sharif University of Technology has been conducting the Atmospheric stability A D

study of the application of quantitative risk assessment Ambient temperature (oC) 27 3
(QRA) on the sitting of compressed natural gas (CNG) Humidity 35% 70%
stations and determining nearby land use limitations [1].
The most important consideration is to be assured that the
proposed site would not be incompatible with existing land Frequency estimation
uses in the vicinity. It is possible by the categorization
of the estimated levels of individual risk (IR) which the Frequency estimation is the methodology used to estimate
proposed site would impose upon them. An analysis of the the number of occurrences of a scenario through a year.
consequences and likelihood of credible accident scenarios Estimates may be obtained from historical incident data on
coupled with acceptable risk criteria is then undertaken. failure frequencies or from failure sequence models, such
The study would determine the safe distance from CNG as FTA [5]. Depending on scenario type both techniques
station borders and how dense the population of the
were used to estimate scenario frequencies as listed in

proposed/existing CNG station. Table 3.
Typical incidents are identified into a few class of risk Table 3. Estimated frequencies of credible scenarios
happening at the CNG station. In Tehran (Iran), they chose
Scenario Estimated
one of the largest CNG stations as their case study to obtain Scenario description
No. Frequency
required information. Low frequency and low consequence
01 Rupture in dryer pipeline 7.5E-5
scenarios were determined to identify the risk estimation.
5mm and 25 mm hole
02, 03 3.8E-5 and 1.0E-7
diameter in cylinders
Once the scenario was defined, source models were selected
to describe how materials are discharged. The source model 5mm hole diameter and
04, 05 0.8E-2 and 1.7E-2
rupture in dispenser pipes
provides a description of the discharge rate and the total
quantity discharged. A dispersion model was subsequently
32
Risk estimation In our present case study, sensitive locations such as

houses, recreational places and high traffic roads are
One popular measure to risk monitoring is IR usually shown located exactly adjacent to CNG station borders which
on a risk contour plot. The IR is defined as the probability are absolutely advised against and these areas must be out
of death at any particular location due to all undesired of the Outer zone characterizing by more than 82 m as
events. The following figure presents the IR contours of a safe distance. Shopping places are also located in close
the selected CNG station. CNG station neighborhood which are advised against too
and they must be out of the Middle zone characterized by
more than 30m as a safe distance.
There are such calculated distances between CNG station

and general acceptable risk borders that usually are not
followed (e.g. present case study), these distances usually
are not intended more important in comparison with other
aspects to determine proper distances such as site area
value and accessibility for vehicles.
Obtained results obviously introduce many limitations to site

CNG station toward following all acceptable risk criteria
for all construction developments, these limitations show
that a large number of parameters should be considered
to select optimal site for a CNG station in a populated
city. This undesirable outcome is almost always present;
to ignore these criteria means imposing unacceptable
risk on people living and working in the neighborhood.
The number of CNG stations and their close vicinity to
populated areas, residential and office buildings and other
IR contonurs for selected CNG Station reasons, especially in Iran, prove that enough studies have
Figure 2. IR Contours for selected CNG Station.
not been done in this field. Although CNG stations have
an important role to play in the country’s economy and
environment, they are hazardous sources; the risks posed
When considering proposals to site a process industry or by CNG stations are poorly evaluated. Further studies on
any development in its neighborhood, four general categories defining acceptable risk evaluation criteria for different
of development are distinguished: industrial, shopping, societies are necessary.
housing and sensitive. Within the Inner zone (where the
IR is greater than 1.0E-5 yr-1) UK HSE normally advises
against all developments other than small or moderate CONCLUSIONS & RECOMMENDATIONS
industrial developments and limited numbers of other
small developments. Within the Outer zone (where the (a) Regular check on the both CNG vehicle and refueling
IR is between 1.0E-6 yr-1 and 3.0E-7 yr-1) only sensitive station must be carried out to meet the standard as
developments are advised against. Across the Middle zone stipulated by the National Fire Protection Association
(1.0E-5 yr-1 to 1.0E-6 yr-1) and where developments or NFPA.
straddle zone boundaries, each development proposal is
considered on its own merits [6, 7]. By comparing these (b) Training on the mitigation measure to deal with
general criteria with numerical results extracted from explosion due to CNG tank or refueling station must
Figure 2, safe distances from CNG station borders can be be conducted. In Singapore, Singapore Civil Defense
determined for each zone as shown in Table 4. Force (SCDF) issues the regulations and mitigation
measures in the contingency of accident related to

Table 4. Safe and real distances from CNG station flammable materials under the Fire Safety Act Chapter
borders for each zone
109A.
Safe distance from Real distance from
station border (m) station border (m) (c) Different type or material of cylinder of CNG will
Inner zone (18) -* affect the safety of the incident rate. The latest material
Middle zone (30) 12 used for the CNG tank (Type 4) and the safest to date
Outer zone (82) 0 is fiber glass hence the cost is expensive.
*No industrial development is available. (d) CNG can help to reduce the reduce greenhouse
gas emissions compare to normal fuel engine system
or diesel.
33
(e) Choosing the site for CNG station needs to consider all [3] CCPS, 2003. “Guidelines for Facility Sitting and Layout”.
the risk criteria to minimize the undesirable outcome. AICHE, New York, USA.
The considerations to site the CNG station to the close [4] CCPS, 2007. “Guidelines for Hazard Evaluation Procedures”.
vicinity are populated areas, residential and office 3rd Edition, AIChE, New York, USA.
buildings. These are highly populated areas where the [5] Harvard Paper Quoted by International at Left, 2003, National
mortal risk is very high. Highway Traffic Safety Administration (DOT) FARS Data
Run for Natural Gas Vehicle Coalition in 2003, USA.
[6] Rosli, A.B., Mohamad F.O. and Semin, A.R.I., 2008. “The
REFERENCES Compressed Natural Gas (CNG) Cylinder Pressure Storage
Technology” in Natural Gas Vehicles (NGV) Research Trends,
[1] Badri, N., Nourai, F. and Rashtchian, D., 2009. “Quantitative USA.
Risk Assessment to Site CNG Refueling stations”. Sharif [7] SAE Paper 2001-01-1343, 2001, “Well-To-Wheel Energy
University of Technology, Iran. Use and Greenhouse Gas Emissions for Various Vehicle
[2] CCPS, 2000. “Guidelines for Chemical Process Quantitative Technologies”. J.J.J. Louis, Shell Global Solutions, USA.
Risk Analysis”. 2nd Edition, AIChE, New York, USA.
34
ADSORPTION THERMODYNAMICS
OF ANTIBIOTICS BY GAC
Shen Liang (shen0042@e.ntu.edu.sg)
Liu Yu (cyliu@ntu.edu.sg)
ABSTRACT: This study investigates the adsorption thermodynamics of three representative β-lactam antibiotics (penicillin G, PCG;
ampicillin, AMP and cephalosporin C, CPC) by GAC. A series of adsorption experiments were carried out at different temperatures.
Results showed that GAC has a substantial adsorptive capacity for the antibiotics studied. Negative ∆G° value suggests that the antibiotic
adsorption by GAC would be spontaneous. ∆H° was estimated as 64.8, 64.4 and 60.3 kJ/mol for PCG, AMP and CPC, respectively,
indicating that the adsorption of three antibiotics by GAC would be endothermic and chemisorption-predominant. FTIR spectra of
antibiotic-loaded GAC further confirmed that antibiotics were adsorbed by GAC through chemical bindings.
INTRODUCTION Langmuir isotherm:

KLCe
Nowadays, antibiotics have emerged in various water bodies qe = qm … (1)
KLCe + 1
due to their expanding production and application in the
world. Antibiotic resistance would thereafter evolve and Freundlich isotherm:
spread in ecosystem, which poses a serious risk for human
health (Kümmerer 2004). Adsorption has been shown to qe = KFC1/n
e … (2)
be effective in removing various soluble organics from
water. Therefore, this study investigates the adsorption in which KL is the Langmuir equilibrium constant (L/mg);
thermodynamics of antibiotic adsorption by granular qe and qm are the equilibrium and maximum adsorption
activated carbon (GAC). capacity, respectively (mg/g); KF is the Freundlich constant;
and n is the heterogeneity factor. Table 1 shows that the
Langmuir isotherm can provide a better description for
MATERIALS AND METHODS adsorption data of PCG, AMP and CPC than the Freundlich
isotherm at all the temperatures studied. The values of
GAC from Calgon Carbon Corporation, USA, was used qm from the Langmuir isotherm indicate that GAC has a
as the adsorbent with a mean size of 2.8 mm, an apparent substantial adsorption capacity for three β-lactam antibiotics
density of 450 kg/m3 and a particle density of 650 kg/m3. studied, and the adsorption capacity of antibiotics by GAC
The GAC was carefully rinsed with distilled water and is in the order of PCG>AMP>CPC.
dried at 103ºC overnight before use. Three representative
β-lactam antibiotics, namely penicillin G (PCG), ampicillin Table 1. Equilibrium constants for the adsorption of
(AMP) and cephalosporin C (CPC) from Sigma-Aldrich antibiotics by GAC.
Pte Ltd, Singapore, were used as the model antibiotics. Freundlich isotherm Langmuir isotherm
Equilibrium experiments were performed with 1.0 to 10.0 Name T/K
KF n R 2
qm KL R2
grams of GAC in 250 mL of the antibiotic solutions with
various concentrations in a temperature controlled shaking PCG 303 261.5 4.44 0.8334 460.2 0.061 0.9711
thermostat (298, 303, 308 and 313K). Concentrations of PCG 308 288.2 4.77 0.7681 485.6 0.081 0.9355
antibiotics were determined by high-performance liquid PCG 313 309.0 4.93 0.6920 502.7 0.102 0.9856
chromatography (HPLC, Perkin Elmer Series 200, USA) AMP 298 65.49 2.24 0.9680 164.2 0.023 0.9981
with a UV detector at 220 nm. Fourier transform infrared

AMP 303 92.32 3.01 0.8777 177.1 0.033 0.9831
(FTIR) spectra of antibiotic-loaded GAC were obtained
from the BioRad Excalibur Series FTS 3000 spectrometer AMP 308 100.7 3.16 0.8320 178.0 0.045 0.9205
(USA). AMP 313 117.4 3.72 0.8410 179.0 0.082 0.9875
CPC 298 33.07 3.28 0.7851 33.67 0.46 0.8543
CPC 303 53.62 3.64 0.7996 50.90 0.49 0.9584
RESULTS AND DISCUSSION
CPC 308 59.14 3.86 0.8460 63.20 0.75 0.9929
Adsorption isotherm CPC 313 67.46 4.03 0.9032 71.9 1.46 0.9879
In this study, the equilibrium data were fitted to the following

isotherm equations,
35
Thermodynamics of adsorption The ∆H° for PCG, AMP and CPC suggests that adsorption
of PCG, AMP and CPC by GAC was a chemical-
The Gibbs free energy indicates the degree of spontaneity dominant process. However, it should be pointed out that
of an adsorption process. The Gibbs free energy change the thermodynamic mechanisms of adsorption would be
(∆G°) of adsorption can be determined as follows: dependent on the chemical structures of various antibiotics.
For example, ∆H° values reported in the adsorption of
∆G° = -RTlnKL … (3) 7-aminocephalosporanic acid, cephalexin, cefadroxyl and
6-aminopenillanic acid by activated carbon ranged from
∆G° is also related to the change in entropy, ∆S° and the 17 to 46 kJ/mol (Dutta et al. 1999), implying a physical-
heat of adsorption, ∆H° at a given temperature in a way chemical mixed mechanism.
such that:
∆G° = ∆H° - T∆S° … (4) FTIR spectrum
Combining Eqs. 3 and 4 yields, The FTIR spectra can be used to identify the functional
groups capable of adsorbing organic compounds. Figure 1
In KL = – ∆H° + ∆S° … (5) shows the FTIR spectra of GAC before and after adsorption
RT R
of antibiotics at 298 K. The common structure of PCG,
Thus, ∆H° and ∆S° can be determined from the slope AMP and CPC is the four-membered (β) lactam ring fused
and the intercept of the linear Van’t Hoff plot, i.e. ln KL to another thiazole ring. Therefore, in FTIR spectra of the
versus 1/T. used GAC, the peak at around 1700 cm-1 represents the
C=O stretching vibration of the β-lactam ring; the peak at
The Langmuir equilibrium constants obtained at 298, 303, around 3480 cm-1 represents the N-H stretching vibration
308 and 313K were used to calculate ∆G°, ∆H° and ∆S° for of the β-lactam ring. Furthermore, the peak at around 650
adsorption of PCG, AMP and CPC by GAC, respectively cm-1 in the spectra of PCG-loaded GAC and AMP-loaded
(Table 2). GAC suggests the C-H bending vibration of the benzene
ring. Meanwhile, this peak did not appear in the spectrum
Table 2. Thermodynamic parameters for the adsorption of CPC-loaded GAC due to the absence of the benzene ring
of antibiotics by GAC. in the CPC molecular structure. The intermediate portion of
the spectrum of 1300-900 cm-1 is often referred to as the
∆H° ∆S° ∆G° (kJ/mol)
Name “fingerprint” region. The absorption pattern in this region
kJ/mol J/mol K 298K 303K 308K 313K is rather complex because the stretching vibration of all
PCG 64.8 295.3 -22.8 -25.2 -26.3 -27.3 single bonds and vibration of molecular skeleton would
AMP 64.4 290.8 -22.4 -23.7 -24.9 -26.9 take place. Compared to the spectrum of virgin GAC,
CPC 60.3 301.9 -30.1 -30.8 -32.4 -34.7
absorbance in the spectra of the antibiotic-loaded GAC
became more intense, showing adsorption of antibiotics
Table 2 shows that ∆H° and ∆S° have positive values, by GAC through chemical bonds.
while negative values for and ∆G°. The positive ∆H°
3488
implies that the adsorption of three antibiotics by GAC d.CPC-loaded GAC
1718
is endothermic. In addition, the value of qm in Table
1 increased with the increase in temperature, which
3480
confirms the endothermic nature of the adsorption of three
antibiotics by GAC. Basically, the heat evolved during the c.AMP-loaded GAC 1718
Absorbance
659
physical adsorption is of the same order of magnitude as
3476
the heat of condensation, i.e., 10 to 20 kJ/mol, whereas 1684
the heat of chemisorption generally falls into the range b.PCG-loaded GAC
647
of 40 to 400 kJ/mol (Bansal and Goyal 2005). Therefore,

it appears that chemisorption would be the predominant
mechanism of these three antibiotics adsorption by GAC.
a. GAC blank
The positive ∆S° suggests the increased randomness at the
solid-solution interface with some structural changes in the
adsorbate/adsorbent and antibiotic affinity to GAC. In fact, 4000 3000 2000
-1
1000
the positive ∆S° is often referred to as an increase in the Wavenumber (cm )
degree of freedom of the adsorbed species. The negative Figure 1. FTIR spectra of antibiotic-loaded GAC at 298 K.
∆G° indicates the degree of spontaneity of the adsorption
process, i.e. the higher negative value of ∆G°, the more
energetically favorable adsorption was. It seems that the
three antibiotics adsorption by GAC would be spontaneous
in the nature (Table 2).
36
CONCLUSIONS REFERENCES
This study demonstrated that β-lactam antibiotics, such as [1] Bansal, R.C. and Goyal, M., 2005. Activated Carbon
PCG, AMP and CPC, can be effectively removed by GAC Adsorption. London: Taylor & Francis.
adsorption. The equilibrium data can be best described [2] Dutta, M., Dutta, N.N. and Bhattacharya, K.G., 1999.
by the Langmuir isotherm. The positive enthalpy value “Aqueous phase adsorption of certain beta-lactam antibiotics
indicates the endothermic nature of the adsorption process, onto polymeric resins and activated carbon”. Separation and
whereas the magnitude of enthalpy suggests that the Purification Technology, 16(3): 213-224.
adsorption of these antibiotics by GAC was chemisorption- [3] Kümmerer, K., 2004. Pharmaceuticals in the environment:
predominant. The FTIR analysis further confirmed that the sources, fate, effects and risks. New York: Springer.
β-lactam ring of these three antibiotics could form the
strong chemical bonds (e.g. C=O and N-H) with GAC’s
functional groups.
37
AN INTERVAL APPROACH FOR

SUPPORTING URBAN WATER
SUPPLY ANALYSIS
X. S. Qin (xsqin@ntu.edu.sg)
Y. Xu (xuye@ntu.edu.sg)
ABSTRACT: An interval-parameter chance-constrained programming (IPSCCP) model is proposed for supporting urban water supply
management under uncertainty. Through incorporating chance-constrained programming (CCP) into an interval linear programming
(ILP) framework, the model effectively deals with uncertainties expressed as not only probability distributions but also as discrete
intervals, and incorporate pre-defined acceptable levels of constraints satisfaction directly into the optimization process. An interactive
two-step sub-modeling method could be used for model solution. The obtained results would be useful for decision makers to gain
an insight into the tradeoffs between environmental and economic objectives and between increased certainties and decreased safeties
(or increased system-failure risks). The study is a new endeavor in advancing an integrated uncertainly-analysis tool for urban water
supply management; the approach could also be applicable to many other water resources management problems.
INTRODUCTION
The shortage of urban water resources has become a major

obstacle for sustainable socio-economic development of
the cities and has aroused much attention over decades.
Integrated Urban Water Supply Management (IUWSM)
focuses on the integrated management of technical aspects
of water services and is effective in relieving the shortage
problems of water resources. However, IUWSM systems
are often complicated with uncertainties. During the past
decades, many inexact optimization techniques were
developed to describe and handle imprecise and uncertain Figure 1. Urban water supply management system.
elements presented in real-world problems. The purpose
of this study is to develop an interval-parameter stochastic Table 1. Parameters related to the water supply system
chance-constrained programming (IPSCCP) model and
Information of Beginning inventory Maximum capacities
apply it to IUWSM system under uncertainties. water sources (×103 m3) (×103 m3)
Dam [15000, 19000] *
(4600, 480) *
Well [2050, 2950] (3800, 365)
URBAN WATER SUPPLY SYSTEM Information of Beginning inventory Maximum capacities
treatment plants (×103 m3) (×103 m3)
In this study, an IUWSM system will be used for Treatment plant 1 [5, 7.5] (1900, 220)
demonstrating the applicability of proposed method. This Treatment plant 2 [10, 13] (3400, 245)
case was adapted from a real case provided by Fattahi & Treatment plant 3 0 +∞
Fayyaz (2010). For many urban areas, it is necessary to Treatment plant 4 0 +∞
develop effective tools for assisting in urban water service Information of Beginning inventory Maximum capacities
providers and government agencies to generate rational reservoirs (×103 m3) (×103 m3)
water resources management scheme. An integrated IUWSM Reservoir 1 [16, 26] (4500, 420)
system, which incorporated water demand management Reservoir 2 [6.5, 13.5] (720, 60)
and water supply system into a general framework, is Reservoir 3 [1, 3.5] (230, 15)
very important for relieving water shortage problems Reservoir 4 [6.5, 13.5] (440, 35)
and realizing balance between water demand and supply. Reservoir 5 [2, 4.5] (230, 15)
Figure 1 shows the structure and components of IUWSM Reservoir 6 [22, 38] (700, 50)
system. The time periods of IUWSM system operation are
Reservoir 7 [4, 6.5] (440, 35)
considered as one year (it has a time interval of month).
Table 1 shows the related parameters within IUWSM Note: [a1, a2] * represents an interval number where a1 and a2 are
the lower and upper bounds, respectively; (m1, d1)** represents a
system, which are assumed as random variables in normal random variable where m1 and d1 are the mean values and standard
distributions and interval numbers, respectively. deviation, respectively.
38
MODEL FORMULATION where f is the net system cost ($); k (k = 1, 2, …, K)

is the index of time periods and K is number of time
For the studied urban water supply system, an interval- periods; j, t, r and z (j = 1, 2, ..., J; t = 1, 2, …, T; r =
parameter stochastic chance-constrained programming 1, 2, …, R; z = 1, 2, …, Z) are indexes of specific water
model can be formulated as follows (Fattahi & Fayyaz sources, treatment plants, reservoirs and consuming zones,
2010): respectively; J, T, R and Z are numbers of water sources,
treatment plants, reservoirs and consuming zones; BJjk is the
recovered water for each water resource j in each month k
(×103 m3); CJTjt, CTRtr and CRZrz are the transferred costs
of water in network, from water sources j to treatments t,
treatments t to reservoirs r and reservoirs r to consuming
zone z, respectively ($); Dzk is the amount of water required
…(1) for consuming zone z in month k (×103 m3); IROr, ITOt
and IJOj are the inventories of each reservoir r, treatment
Subject to: t and water resource j at the first of the planning horizon,
respectively (×103 m3); IRrk, ITtk and IJjk are the inventories
…(2) of each reservoir r, treatment t and water resource j at
end of each month k, respectively (×103 m3); LXJjt, LXTtr
and LXZrz are the leakage rates of water in network, from
…(3) water sources j to treatments t, treatments t to reservoirs
r and reservoirs r to consuming zone z, respectively (%);
MJjk is the maximum amount of water that can be exited
from water sources j at each month k (×103 m3); PRjk is
the purchasing cost of water from water sources j at each
…(4)
month k ($); qz is the acceptable level of constraints-
…(5) satisfaction. TL is the allowed maximum leakage amounts
(×103 m3); VRrk and VTtk are the capacities of reservoirs r
…(6) and treatment t at each month k (×103 m3); XJTjtk, XTRtrk
and XRZrzk are decision variables, representing the amount
…(7) of water transferred from water sources j to treatments
t, from treatments t to reservoirs r and from reservoirs
r to consuming zones z at each month k, respectively
(×103 m3); ZRZrz, ZJTjt and ZTRtr are binary variables (i.e.
…(8) expressed as 1 or 0, representing yes or no answers) used
to define paths in the network, respectively. Referring
…(9) to the proposed model, the constraints with interval and
random coefficients (i.e. constraints 2, 7, 11 and 14) can
…(10) be transformed to their respective crisp equivalent (Charnes
et al. 1972; Huang et al. 1992; Qin et al. 2007). Referring
…(11) to the proposed model, the transformed ILP models can be
formulated and solved, such that the objective values and
decision variables expressed as discrete intervals at various
…(12)
constraints-violation levels can be obtained.
…(13)
…(14) The solutions at a number of acceptable probability levels

(i.e. 0.9, 0.95 and 0.99) are obtained through solving the
IPSCCP model. The results indicate that the water supply
…(15) patterns based on the demand amounts would be affected
by multiple factors. Firstly, the objective function value and
part of the decision variables from IPSCCP would present
as discrete intervals rather than fixed values. For example,
…(16) at a significance level of 0.9, the objective function value
(i.e. total system cost) would range from 38.68 to 74.26
39
(×106) dollars. The lower bound of the objective function

represents an optimal decision scheme with the lowest cost;
correspondingly, the obtained decision variables would reach
their lower bounds. Conversely, the solution corresponding
to the higher bound of system cost is of conservative
consideration. Based on the obtained interval solutions, a
variety of alternatives can be generated through adjusting
within their solution intervals.
Figure 2 indicates that the variations in the acceptable Figure 3. Comparison of solutions between
IPSCCP and SCCP models.
levels would result in changes of water supply patterns.
For example, in the entire planning period, the total water
amounts transferred from water sources to treatment CONCLUSIONS
plants at significance levels of 0.99, 0.95 and 0.99 would
be [38,287.11, 52,396.04], [38,900.23, 53,224.34] and An interval-parameter stochastic chance-constrained
[40,050.34, 54,778.08] (×103 m3), respectively. The amounts programming (IPSCCP) model was proposed for urban
from reservoirs to consuming zones are [37,305.83, water supply management. IPSCCP could effectively deal
48,760.20], [37,895.13, 49,511.91] and [39,000.55, with uncertainties expressed as both discrete intervals and
50,922.00] (×103 m3), respectively. The reason is that, as random variables. It was also capable of incorporating a set
the increases of the acceptable levels, the constraints would of pre-defined acceptable levels of constraint satisfaction into
become stricter. optimization process, allowing model solutions to achieve
higher system costs at allowable violation probabilities. The
Figure 3 presents the variation of system cost at various results indicate that IPSCCP could help decision makers
acceptable levels. Generally, the system cost would increase gain in-depth insights into the trade-offs between increased
as the increase of acceptable levels. For example, at different system benefits and decreased safeties, and establish rational
significance levels (from 0.90 to 0.99), the system costs are water supply patterns under complex uncertainties for
[38.68, 74.26], [39.19, 75.27] and [40.15, 77.17] (×106 $), meeting the city’s water demand. The method could also
respectively. This is because, as the acceptable level goes be applicable to many other environmental problems.
higher, the constraints would become stricter. To compare the
proposed IPSCCP model with other alternatives, a general
SCCP model is formulated for the same problem where REFERENCES
the deterministic parameters are derived by averaging the
upper and lower bounds of intervals from IPSCCP model. [1] Charnes, A., Cooper, W.W. and Kirby, P., 1972. “Chance
As shown in Figure 2, the total water amounts transferred constrained programming: An extension of statistical method”.
from water sources to treatment plants at significance levels New York: Optimizing Methods in Statistics, Academic
Press.
of 0.99, 0.95 and 0.99 would be 44,308.26, 45,013.78
and 46,337.20 (×103 m3), respectively. The amounts from [2] Fattahi, P. and Fayyaz, S., 2010. “A compromise programming
model to integrated urban water management”. Water
reservoirs to consuming zones are 42,198.95, 42,858.26 and
Resources Management, 24: 1211-1227.
44,095.02 (×103 m3), respectively. In reference to Figure
[3] Huang, G.H., Baetz B.W. and Patry, G.G., 1992. “A grey linear
3, the total cost at different significance levels are 70.88,
programming approach for municipal solid waste management
71.88 and 73.76 (×106 $), respectively. In such a case, the
planning under uncertainty”. Civil Engineering Systems, 9:
decision alternative would be restricted to a single solution, 319-335.
which may limit its application in real-world systems.
[4] Qin, X.S., Huang, G.H., Zeng, G.M., Chakma, A. and
Huang, Y.F., 2007. “An interval-parameter fuzzy nonlinear

optimization model for stream water quality management
under uncertainty”. European Journal of Operational Research,
180(3): 1331-1357.
Figure 2. Total water amounts transferred in the network.
40
ANAEROBIC HYDROLYSIS OF
PARTICULATES IN SEWAGE
Teo Chee Wee (teoc0044@ntu.edu.sg)
Philip Wong (pcywong@ntu.edu.sg)
ABSTRACT: Anaerobic hydrolysis of particulates presents both a challenge and an opportunity for the treatment of dilute wastewaters such
as sewage. A significant portion of the total chemical oxygen demand of raw sewage is in the form of particulates. These particulates have
to be hydrolyzed prior to assimilation by the anaerobic consortia. Through enhancing hydrolysis kinetics, the overall biotransformation
rate can be increased, chemical energy in the organic solids can be efficiently tapped, and solids accumulation can be better managed.
This study investigates the optimum conditions for anaerobic hydrolysis of particulates in sewage treatment, with a focus on enzyme
augmentation with hydrolases. The operating conditions investigated were pH, temperature and enzymes dosage. Experimental results
showed a higher degree of hydrolysis at pH 5.0 showed higher degree of hydrolysis compared to those at pH 9.0. In addition, control
experiments revealed considerable degradation of enzymes in the batch reactors, which may be attributed to proteolysis of amylases
and autolysis of proteases.
INTRODUCTION bacteria produce exo-enzymes which break up polymers

forming short chain dimers and monomers, which are in
Hydrolysis in anaerobic sewage treatment turn further degraded by other microbes for catabolism and
anabolism. The polymers have to be hydrolyzed to low
One feature of anaerobic digestion in organic waste molecular weight monomers (≤ 1000 Dalton) before they
stabilization is the concomitant production of energy in the can be assimilated into cells (Burgess et al., 2008).
form of methane and hydrogen gas. This technology can
potentially be exploited to recover valuable energy directly Complex wastewater such as sewage contains carbohydrates,
from raw sewage. In anaerobic digestion, hydrolysis of lipids and proteins. Carbohydrates are known to be
suspended solids tends to be rate limiting especially for hydrolyzed rapidly to simple sugars (Zeeman & Sanders,
substrates with high solids content. This is also the case for 2001). Domestic wastewater typically comprises of 40-60
treatment of domestic wastewater (Seghezzo et al., 2005). percent of proteins, 25-50 percent of carbohydrates and 8-12
Hydrolysis is also vital for the degradation of dissolved percent of oil and fats (Asano et al., 2007). Additionally,
macromolecules and soluble microbial products (SMPs) urea and some synthetic organic chemicals may also be
that are prevalent in wastewater. present. Proteins are hydrolyzed into amino acids, lipids are
hydrolyzed into long chain fatty acids (LCFAs) and glycerol,
Domestic wastewater contains approximately 500-800mg/l and carbohydrates are hydrolyzed into simple sugars.
of total solids with a suspended solids concentration of
approximately 155-330mg/l (US EPA, 2008). In terms Anaerobic hydrolysis of complex substrates can proceed in
of chemical oxygen demand (COD), suspended solids a number of ways (Angelidaki et al., 2004). For instance,
represent up to 85% of the total COD (Tarek et al., 2001). the hydrolytic bacteria can secrete exo-enzymes into the
Hence, hydrolysis of organic solids in sewage represents bulk solution to hydrolyze the complex substrate. Hydrolytic
an important step in reducing the effluent COD as well bacteria can also attach itself to the particulate and excrete
as to improve overall treatment efficiency. Hydrolysis of exo-enzymes to degrade it. Finally, the hydrolytic bacteria
organic solids also produces short chain fatty acids that can adsorb to the particulate surface and utilize their attached
can be utilized for downstream biological nutrient removal enzymes (ectoenzymes) to initiate hydrolysis.
(Feng et al., 2009).
Enhancing hydrolysis of particulates

Hydrolysis of particulate macromolecules
Several pretreatment techniques had previously been
Hydrolysis is a reaction in which the chemical bonds investigated. Tanaka et al. (1997) studied the effect of
in polymers are cleaved by the introduction of water thermochemical pretreatment on the anaerobic digestion
molecules. H+ ion is added to a fragment of the polymer of waste activated sludge. The pretreatment solubilized the
whereas OH- ion is added to the other fragment from the volatile suspended solids (VSS) by 40-50% and increased
same polymer. From a biochemical perspective, hydrolytic methane production by more than 200% relative to the
41
control without pretreatment. Neis et al. (2000) studied Seed sludge was obtained from the anaerobic digester at
the enhancement of sludge hydrolysis by ultrasonic Ulu Pandan Water Reclamation Plant, Singapore. The seed
pretreatment. The process improved sludge degradation rate sludge was washed twice with phosphate buffered saline
by 30% at a solids residence time (SRT) of 16 days. (PBS) to remove residual soluble COD. Sludge was added
to the batch reactors such that the starting concentration
was 8000mg MLSS/l (5000mg MLVS/l).
Enzymatic hydrolysis as pretreatment of lipids-rich
wastewater using low cost lipase prepared from porcine
The exogenous enzymes added were liquid enzymes blend
pancreas was investigated by Adriano et al. (2006). The (BioCat Microbials Pte Ltd, USA) containing proteases,
wastewater was hydrolyzed prior to addition into the lipases and amylases. The blend was purified in laboratory
bioreactor. Higher COD removal (69-80%) and biogas using a stirred cell with 10kDa ultrafiltration membrane
production (10-12ml/g COD) compared to the control disc (Amicon PM 10, Millipore Co., USA) to remove
without enzymatic pretreatment (COD removal of 40% the low molecular weight propylene glycol before each
and biogas production of 6ml/g COD). Moreover, it was experiment.
found that direct addition of enzymes into the bioreactor is
viable and attractive, with high COD removal of 76.4%. The Erlenmeyer flasks (0.5L) with rubber stoppers were used
feasibility of enzymatic hydrolysis was also demonstrated as batch reactors. The headspace of the flasks was flushed
by others authors (Leal et al., 2006; Rosa et al., 2009; with nitrogen gas before sealing with rubber stoppers.
A water bath shaker was used to provide the necessary
Lee et al., 2008).
temperature and agitation. pHs were adjusted daily with
6N HCl and 6N NaOH.
Hydrolysis of particulates can be enhanced by operating at
the optimal pH and temperature. The rate is higher under TSS, VSS and sCOD were quantified according to the
thermophilic condition compared to mesophilic condition. Standard Methods (APHA, 1998). DOC was analyzed with
Within each range, hydrolysis rate proceeds faster at higher the TOC ASI-V (Shimadzu Co., Japan) TOC analyzer.
temperatures. In general, the optimum pH is at near neutral Samples for sCOD and DOC assays were filtered with
(Rollon, 1999), although Isaacson (1990) pointed out that, 0.45μm syringe filters. Protease assay was performed
neutral pH commonly operated in anaerobic digesters aims using Sigma Aldrich assay kit with casein as the substrate
to maximize the rate of methanogenesis; hydrolysis may (substrate product no.: C7078). α-amylase assay was
have a different optimum. For example, the optimum pH performed using the dinitrosalicylic acid method (Bernfeld,
of holocellulytic bacteria is species dependent and can 1955). Biogas production rate was measured by the syringe
displacement method (Owen et al., 1979). One unit of
range from 3 to 11. The optimum pH for hydrolysis and
protease activity is defined as the amount of enzyme that
acidification is 5.6-6.0 whereas that for methanogenesis is
hydrolyzes casein to produce 1.0μmol of tyrosine per minute
6.8-7.2 (Riva, 1992). Therefore it can be inferred that the at pH 7.5 and 35oC. One unit of amylase activity is defined
optimum pH for hydrolysis varies according to the substrate as the amount of enzyme that hydrolyzes starch to liberate
composition and the hydrolytic species present. 1mg of maltose in 3 minutes at pH 7 and 25oC.
The objective of this research is to investigate the optimum The experiments were performed in batches of 20 days.
conditions for the anaerobic hydrolysis of particulates in The degree of hydrolysis of particulates was calculated
sewage including a preliminary study on the feasibility of from the following equation by analyzing the period where
enzyme augmentation. hydrolysis predominates:
Degree of hydrolysis = [(∆sCOD + ∆CODH2 + ∆CODCH4)

/ (Initial particulate COD)] (1)
MATERIALS AND METHODS
where ∆CODH2 is the sCOD converted into biogas as H2
The synthetic wastewater used in the experiments was (8g COD/g H2); ∆CODCH4 is the sCOD converted into
prepared to simulate raw municipal wastewater with biogas as CH4 (4g COD/g CH4). Hence, the degree of
suspended solids. The soluble constituents of the wastewater hydrolysis is expressed in grams of COD hydrolyzed per
were prepared according to the OECD guideline for synthetic gram of particulate COD.
sewage (OECD 303A). The twice diluted composition gives
a COD of 150mg/l. Suspended solids in wastewater are Temperature and pH were investigated in the range of 25-
simulated by 250mg/l of dry dog food with a COD of 35oC and pH 5.0-9.0 respectively. Enzymes were augmented
400±50mg/l (ALPO, Purina Co., USA). ALPO dog food at the concentration of 0-0.7% (w/v).
was selected because it has similar organic composition
with primary sludge (Kim et al., 2003). This gave a total
COD of 550±50mg/l.
42
Degree of hydrolysis
The degree of hydrolysis is calculated from the period where

hydrolysis predominates, as indicated by rising sCOD. The
calculated degree of hydrolysis at these periods (indicated in
parentheses in days) and the total degree of hydrolysis for
the respective batch experiments are shown in Table 1.
Complete degradation of initial sCOD by Day 1 is assumed

as it is easily degradable and low in concentration (COD of Figure 1: Control experiment.
100mg/l). For computation that includes Day 0-1, 100mg/l
is deducted from the COD balance to ensure accountability. low precision of VSS test as no replicates are performed.
At pH 5.0 and 9.0, COD removal due to biogas production Subsequently, there is a decrease in sCOD of 170mg/l
is assumed to be negligible as methane and hydrogen from Day 1-7, suggesting the consumption of enzymes as
production is low at these pH [Lay et al., 1997; Wang substrate by the biomass. One possible mechanism is the
proteolysis of lipases and amylases as well as the autolysis
et al., 2007; Lin et al., 2006]. The majority of the COD
of proteases into monomers that can be consumed by
remains in the aqueous phase as soluble compounds like
the microorganisms. In addition, hydrolytic bacteria may
volatile fatty acids (VFAs), long chain fatty acids (LCFAs) secrete hydrolases to degrade the enzymes. There is a rise
and alcohols. in sCOD from Day 7 to 13 (69 to 143mg/l), suggesting
the production of SMPs due to cell death and lysis. The
Table 1: Degree of hydrolysis at different conditions fall in MLVSS from 4850 to 4350mg/l in the same period
Conditions Period 1 Period 2 Period 3 Total indicates that the microorganisms are probably undergoing
endogenous respiration.
pH 9, 35 C,
o
0.415 0.1725 - 0.5875
0.7% (w/v) (3-9) (15-20)
Control without the addition of biomass was performed in a
pH 9, 25oC, 0.2775 0.0325 0.145 0.455 sterilized 250ml Duran bottle for a period of 4 days. Gradual
0.7% (w/v) (2-6) (8-10) (16-20) decline in enzymatic activity was observed throughout
pH 5, 25oC, 0.2025 0.105 0.35 0.6575 the experimental period, suggesting the occurrence of
0.7% (w/v) (0-1) (7-9) (10-13) proteolysis and autolysis. Protease activity fell from
pH 5, 35oC, 0.9425 0.18 0.225 >1 1.05U/ml to 0.22U/ml whereas amylase activity fell from
0.7% (w/v) (0-1) (7-8) (15-17) 2.15U/ml to 0.079U/ml during the same period.
In the context of enzyme augmentation, it is possible for the

calculated degree of hydrolysis to exceed 1 as the enzymes CONCLUSIONS
can exert COD. From the experimental results, it can be
deduced that lower pH enhances the rate of anaerobic Preliminary results indicate that lowering pH can enhance
hydrolysis of particulates in municipal wastewater. Other the rate of hydrolysis of particulates in sewage. The
authors reported an optimum pH of 5.5 for hydrolysis of decline in enzymatic activity over time indicates that free
diary wastewater (Yu et al., 2002), 6.5 for hydrolysis of enzymes may not be stable in continuous operation like
synthetic PS (Kim et al., 2003) and 7.0 for hydrolysis of anaerobic membrane bioreactor. Future works involve the
kitchen waste (Zhang et al., 2005). study of optimum rate of hydrolysis using response surface
methodology and verification in laboratory scale anaerobic
membrane bioreactor.
Control with enzymes

REFERENCES
A control was performed with only enzymes and no
substrate added. Tap water was added as the source of [1] APHA, 1998. “Standard methods for the examination of water
micronutrients in place of synthetic feed. The result of the and wastewater”. American Public Health Association, 20th
control is illustrated in Figure 1. Edition.
[2] Asano and Takashi, 2007. “Water reuse: Issues, technology
The data points are connected with straight lines for better and applications”. McGraw-Hill publication, pp. 34-58.
visualization. There is a rise in sCOD of 31mg/l from [3] Adriano A. Mendes, Ernandes B. Pereira and Heizir F. de
Day 0-1, suggesting the hydrolysis of biomass. This may Castro, 2006. “Effect of the enzymatic hydrolysis pretreatment
be corroborated by MLVSS, which falls from 5150 to on lipids-rich wastewater on the anaerobic digestion”.
5000mg/l. The low COD per unit of VSS may be due to the Biochemical Engineering Journal, 32(2006), pp. 185-190.
43
[4] Analiza Palenzuela Rollon, 1999. “Anaerobic digestion of fish [15] Jo E. Burgess and Brett I. Pletschke, 2008. “Hydrolytic
processing wastewater with special emphasis on hydrolysis enzymes in sewage treatment: A mini review”. Water SA,
of suspended solids”. PhD thesis, Wageningen Agricultural Vol. 34, No. 3, pp. 343-350.
University, pp. 21-26. [16] Jiunn-Jyi Lay, Yu-You Li and Tatsuya Noike, 1997. “Influence
[5] Bernfeld O., 1955. “Amylases, alpha and beta”. Methods of pH and moisture content on the methane production in
in Enzymology, 1(1955), pp. 149-158. high solids sludge digestion”. Water Research, Vol. 31, No.
[6] B. Zhang, L-L Zhang, S-C Zhang. H-Z Shi and W-M Cai, 6, pp. 1518-1524.
2005. “The influence of pH on hydrolysis and acidification [17] Marcia C.M.R. Leal, Denis M.G. Freire, Magali C. Cammarota
of kitchen wastes in two phase anaerobic digestion”. and Geraldo L. Sant’ Anna Jr, 2006. “Effect of enzymatic
Environmental Technology, 26(2005), pp. 329-339. hydrolysis on anaerobic treatment of dairy wastewater”.
[7] Chiu-Yue Lin, Chun-Hsiung Hung, Chau-Huei Chen, Wan- Process Biochemistry, 41(2006), pp. 1173-1178.
Ting Chung and Lu-Hsiu Cheng, 2006. “Effects of initial [18] M. Kim, C.Y. Gomec, Y. Ahn and R.E. Speece, 2003.
cultivation pH on fermentative hydrogen production from “Hydrolysis and acidogenesis of particulate organic material
xylose using natural mixed cultures”. Process Biochemistry, in mesophilc and thermophilic anaerobic digestion”.
41(2006), pp. 1383-1390. Environmental Technology, Vol. 24, pp. 1183-1190.
[8] Ching-Hsiung Wang, Wei-Bin Lu and Jo-Shu Chang, 2007. [19] Ron Isaacson, 1990. “Methane from community wastes”.
“Feasibility study on fermentative conversion of raw and Tylor and Francis group, pp. 77-86.
hydrolyzed starch to hydrogen using anaerobic mixed [20] Seghezzo L., Zeeman G. and Lettinga G., 2005. “Anaerobic
microflora”. International Journal of Hydrogen Energy, sewage treatment”. Water Encyclopedia, pp 517-521.
32(2007), pp. 3849-3859.
[21] Shuzo Tanaka, Toshio Kobayashi, Kenichi Kamiyama, Ma
[9] Daniela R. Rosa, Iolanda C.S. Duarte, N. Katia Saavedra, Lolita and N. Signey Bildan, 1997. “Effect of thermochemical
Maria B. Varesche, Marcelo Zaiat, Magali C. Cammarota pretreatment on the anaerobic digestion of waste activated
and Denise M.G. Freire, 2009. “Performance and molecular sludge”. Water Science and Technology, Vol. 35, No. 8, pp.
evaluation of an anaerobic system with suspended biomass 209-215.
for treating wastewater with high fat content after enzymatic
[22] Tarek A.E., Jesko S., Arie D.K., Harry B., Grietje Z. and
hydrolysis”. Bioresource Technology, 100 (2009), pp. 6170-
Gatze L., 2001. “Biodegradability and change of physical
6176.
characteristics of particles during anaerobic digestion of
[10] Feng Leiyu, Yan Yuanyuan and Chen Yinguang, 2009. domestic sewage”. Water Research, Vol. 35, No. 5, pp.
“Kinetic analysis of waste activated sludge hydrolysis and 1311-1317.
short chain fatty acid production at pH 10”. Journal of
[23] US EPA, 2008. “Chapter 3: Establishing treatment system
Environmental Sciences, 21(2009), pp 589-594.
performance requirements”. EPA 625/R-00/008-Chapter 3.
[11] G. Zeeman and W. Sanders, 2001. “Potential of anaerobic
[24] U. Neis, K. Nickel and A. Tiehm, 2000. “Enhancement of
digestion of complex waste(water)”. Water Science and
anaerobic sludge digestion by ultrasonic disintegration”. Water
Technology, Vol. 44, No. 8, pp. 115-122.
Science and Technology, Vol. 42, No. 9, pp 73-80.
[12] G. Riva, 1992. “Utilization of renewable energy source and
[25] W.F. Owen, D.C. Stuckey, J.B. Healy, JR. L.Y. Young,
energy saving technologies by small scale milk plants and
P.L. McCarty, 1979. “Bioassay for monitoring biochemical
collection centres”. Food and Agricultural Organization, pp.
methane potential and anaerobic toxicity”. Water Research,
39-40.
13(1979), pp. 485-492.
[13] H-Q Yu and H.H.P. Fang, 2002. “Acidogenesis of dairy
[26] Y.H. Lee, Y.C. Chung and J.Y. Jung, 2008. “Effect of
wastewater at various pH levels”. Water Science and
chemical and enzymatic treatments on the hydrolysis of
swine wastewater”. Water Science and Technology, Vol. 58,
[14] Irini Angelidaki and Wendy Sanders, 2004. “Assessment of No. 7, pp. 1529-1534.
the anaerobic biodegradability of macropollutants”. Reviews
in Environmental Science and Biotechnology, 3(2004), pp.
117-129.
44
DATA-DRIVEN APPROACH FOR MULTI-

STEP AHEAD FLOOD FORECASTING FOR
THE LOWER MEKONG
Nguyen Khac-Tien Phuoc (nguy0077@ntu.edu.sg)
Chua Hock-Chye Lloyd (chcchua@ntu.edu.sg)
ABSTRACT: Accurate flood forecasts are essential for an early flood warning system to issue reliable flood warnings in order that proper
emergency actions be taken to mitigate flood damages. This study proposes a data-driven approach based on the Adaptive-Network-
Based Fuzzy Inference System (ANFIS) to forecast the water level for the Lower Mekong River at Pakse in Laos for lead-times from 1
to 5 days. In addition, the present study used an output updating scheme which is based on a recursive autoregressive (RAR) model to
enhance the accuracy of multi-step ahead forecasts. The results of the present study show that: (i) ANFIS model performed well for 1-,
2- and 3-days ahead forecasts when compared to the performance of a physically based model which is currently adopted for operational
forecasts; and (ii) the output updating technique significantly enhances 4- and 5-days ahead forecasts.
INTRODUCTION An alternative flood forecasting tool, based on the data-

driven approach, which utilizes the Adaptive-Network-
The Mekong River with a length of approximately 4,800 Based Fuzzy Inference System or ANFIS (Jang, 1993) was
km has its source in China’s Yunnan province and flows developed for this study. ANFIS focuses on constructing
through Myanmar, Thailand, Laos, Cambodia and Vietnam an input-output mapping based on the measured time
and discharges into the South China Sea. It drains an area series data. Studies where ANFIS has been used in flood
of approximately 795,000 km2 and has a yearly average forecasting include Nayak et al. (2005) and Chen et al.
flow rate of 15,000 m3/s (MRC, 2005). The wet season in (2006).
the Mekong occurs from the months of June to October
every year. Floods may occur during this time when This study demonstrates the use of ANFIS incorporating
typhoons from the South China Sea cause heavy rains over a recursive autoregressive or RAR based output updating
the basin. In order to deal with flood management and technique for multi-step flood forecasting for a reach of
mitigation in this region, an effective flood early warning the Lower Mekong River. The performance of the ANFIS
system is essential. model was compared with the benchmark used by the
MRC to evaluate forecast accuracy and a Naïve model
Before 2009, flood forecasting for the Lower Mekong was which is the simplest forecast model that is often used as
managed by a system that utilized models and concepts base line to evaluate models. Finally, ANFIS model results
developed during the 1970s (Apirumanekul, 2006). The were also compared with results obtained from the URBS
Streamflow Synthesis and Reservoir Regulation model model for 2009 forecasts.
(SSARR), developed by the US Corps of Engineers was
used to model the rainfall-runoff process for the upper
reaches from Chiang Saen to Pakse, while regression METHODS
models were used to model the flows downstream from
Pakse. A new flood forecasting system has recently been Adaptive-Network-Based Fuzzy Inference System
developed and put into operation at the Regional Flood (ANFIS)
Management and Mitigation Centre of the Mekong River
Commission (RFMMC-MRC). This system was developed A neuro-fuzzy system is a hybrid system combining the
in 2007 in order to improve short and especially medium concepts of Fuzzy Inference System (FIS) and Artificial
term forecasts. The rainfall-runoff processes for the upper Neural Network (ANN). A common framework of a
reach from Chiang Saen to Pakse is now modeled by the neuro-fuzzy system is to represent the FIS in an ANN
Unified Run-off Basin Simulation (URBS) hydrological architecture and utilize the learning capability of ANN to
model, and the flow routing downstream from Pakse is optimize the FIS parameters. The Adaptive Network-based
modeled by the ISIS hydrodynamic model. The system was Fuzzy Inference System (ANFIS) proposed by Jang (1993)
operated in “test mode” during the 2008 flood season, and is a well-know approach for neuro-fuzzy systems. ANFIS
has been in operational mode since 2009 (MRC, 2009). is based on the Tagaki-Sugeno-Kang (Takagi and Sugeno,
45
1985) fuzzy inference system embedded within the structure Recursive Autoregressive Model
of the ANN. Figure 1 shows a schematic of ANFIS with 2
rules for a system consisting of 2 inputs and 1 output. In order to improve the forecast accuracy, the present
study modelled the time-series of water level forecast
errors. Once the forecast error was estimated, it could then
update the forecast from the ANFIS model. Auto-regressive
(AR) model is a stochastic model that is typically used
for this purpose (Serban and Askew, 1991; WMO, 1992).
However, this is a non-adaptive model since its parameters
are fixed after calibration. Therefore it is not able to adapt
to changes once the model is calibrated. The recursive
autoregressive (RAR) model is an adaptive version of the
Figure 1. Schematic of ANFIS architecture (Jang, 1993).
linear autoregressive model. Its parameters can be adjusted
by a recursive estimation algorithm using the most recent
error. Hence, the RAR model is more suitable compared
The function of each layer in ANFIS is briefly described to the AR model in online applications. The present study
as follows: used an RAR model with a recursive estimation algorithm
Layer 1 - input nodes: each node generates membership (Ljung, 1999) implemented in the System Identification
grades to which inputs at the inputs nodes are assigned, from Toolbox of MATLAB (MATLAB, 2008).
fuzzy sets based on the membership functions used. In our
study the Gaussian membership function was applied.
Layer 2 - rule nodes: in this layer, the AND or the OR CASE STUDY
operator is applied to obtain one output that represents
the result of the antecedent for that rule, i.e. the firing The most upstream station of the Lower Mekong is at
strength. Chiang Saen in Laos and the most downstream station
is at Chau Doc in Vietnam. The upstream reach of the
Layer 3 - average nodes: in this layer, the main objective
Lower Mekong from Chiang Sean to Pakse is characterized
is to calculate the ratio of the ith rule’s firing strength to
by steeper gradients compared to the lower reach where
the sum of all rules’ firing strength.
gradients are generally flatter. The URBS model, which is
Layer 4 - consequent nodes: in this layer, the first-order a physically-based lumped parameter model, is currently
Sugeno fuzzy model is adopted. The fourth layer computes used to model the flow upstream of Pakse. Because of the
the contribution of each rule towards the total output. flatter terrain, a 1-D hydrodynamic model (ISIS) is used
Layer 5 - output nodes: the single node computes the overall to model the flow downstream of Pakse. Thus, Pakse is
output by summing all the incoming signals according to the boundary between the URBS and ISIS models and the
the defuzzification process, where each rule’s fuzzy results forecast at Pakse is used as input or upstream boundary
are transformed into a crisp output. condition for ISIS. Therefore, the accuracy of forecasts
at Pakse is important as it can significantly influence the
The learning algorithm in ANFIS optimizes the ANFIS accuracy of downstream forecasts.
parameters that include the premise parameters, which
describe the shape of the membership function, and the
consequent parameters, which describe the overall output ANFIS MODEL DEVELOPMENT
of the system. The algorithm used is a hybrid learning
algorithm consisting of the gradient descent and least- Data division
squares methods. The gradient descent method is employed

to tune the premise parameters, whereas the least-squares The water level of the Mekong at Pakse and upstream
method is used to identify the consequent parameters. The stations in the wet seasons from 1993 to 1998 were used
present study used ANFIS implemented in the Fuzzy Logic for training, while data from 1999 to 2000 were used for
Toolbox of MATLAB (MATLAB, 2008) to forecast water testing, and data from 2001 to 2003 and 2009 were used
level of the Lower Mekong at Pakse in Laos. for validation. ANFIS model results are compared to the
URBS results for the wet season in 2009 only, since URBS
model results are available only for that year. Table 1 shows
some pertinent information of water levels at Pakse in the
three data subsets.
46
Table 1. Statistical properties of training, test, a membership function (e.g., Gaussian, Triangular) is defined
and validation data sets. with the center located at the center of the class. There
Mean Duration Min Max are two methods to generate an initial model for ANFIS
(m) (days) (m) (m) training which are based on two partitioning techniques
on the data: grid partition and subtractive clustering. It
Training
5.98 1,134 1.36 13.01 is not advisable to use grid partitioning in ANFIS when
(1993-1998)
the input dimension is more than five or six (Nayak et
Test al., 2005) due to excessive propagation of the number of
7.12 380 2.56 13.34
(1999-2000) rules (curse of dimensionality). The present study used
Validation the subtractive clustering technique to generate the initial
(2001-2003 and 6.88 693 1.45 12.7 model for ANFIS training. A proper number of membership
2009) functions applied to each input was identified by trial-and-
error. The number of membership functions was increased
from three to twenty. It was found that when the number
Input-output selection of membership functions increased from three to ten the
performance of the model was more or less the same on
ANFIS is a Multi-Input-Single-Output (MISO) system. the test data set, and when the number of rules was greater
Each ANFIS model has only one output. Therefore, five than ten, model performance deteriorated. The number of
ANFIS models providing outputs of L-days (L = 1, 2, … membership functions used was thus three. This would
, 5 days) ahead forecasts of the water level at Pakse were imply that the range of water level can be divided into
developed. This study employed a statistical approach three main regions: low, medium and high.
suggested by Sudheer et al. (2002) to identify the appropriate
input vector. The method is based on the heuristic that the
potential influencing variables corresponding to different OUTPUT UPDATING
time lags can be identified through statistical analysis of
the data series using cross correlation, autocorrelation, and Five independent ANFIS models were developed to
partial autocorrelation between the variables in question. provide multi-step ahead water level forecasts at Pakse.
Correlation analyses showed that the water level at Pakse The Lth (L = 1, 2, … , 5 days) ANFIS model is denoted
at L-days ahead is most related to three recent water levels by ANFISL. Its output predicted at time t is denoted by
at Pakse, then followed by the three recent water levels at (L). This is the forecast made at time t for the water
Savanakhet, which is situated 240 km upstream of Pakse. level at the lead time t+L. The difference between the
Thus, the general form of the relationship between the actual measurement Ht+L, to be measured at time t+L, and
water level at Pakse at step t+L and water levels at Pakse (L) is the output error of the ANFISL model predicted at
and Savannakhet can be expressed as follows: time t. This error can be stated as:
et(L) = Ht+L – (L)
This means that each ANFIS model has as output, the At time t, et(L) is unknown since Ht+L is unknown. An
L-days ahead forecast of the water level at Pakse and the output updating procedure is implemented which attempts
three most recent water levels at Pakse and Savanakhet to approximate the value of et(L) denoted by . Once
as inputs. is determined, the predicted error can then be added
to the original output (L) in order to obtain the updated
output, , as follows:
Model evaluation
= (L) +
The coefficient of efficiency (COE) was used to assess the
overall goodness of fit. The mean absolute error (MAE), An incremental updating procedure was used in the present
mean percentage absolute error (MPAE), and root mean study that utilised output of ANFIS1 to update the output
square error (RMSE) were adopted as absolute error of ANFIS2, the outputs of ANFIS1 and ANFIS2 to update
measures. In addition, the mean absolute error of forecasts the output of ANFIS3 and so on. Figure 2 shows that the
for water levels higher than the alarm water level (MAEhigh difference between output of ANFISL at time step t and the
= 11 m) at Pakse was used to assess model performance output of the ANFISL+1 at time step t-1 is used to estimate
at high river stages. error of the ANFISL+1 output at time step t-1. This estimated
error is applied to an RAR model to obtain the estimated
error of the ANFISL+1 output at time step t. This error is
Model structure selection and model calibration then added to the original ANFISL+1 output to obtain the
updated forecast.
In ANFIS, each input variable is clustered into several class
values in layer 1 to build up the fuzzy rules. In each class,
47
1.25
Benchmark
1.00 Naïve
ANFIS
MAE (m)
0.75
0.50
0.25
0.00
(a) 1 2 3 4 5
Lead-time (day)
1.25
Benchmark
1.00 Naïve
MAE high (m)

ANFIS
0.75
0.50
0.25
Figure 2. Incremental output updating procedure
based on the I-RAR algorithm. 0.00
(b) 1 2 3 4 5
Lead-time (day)
RESULTS AND DISCUSSIONS Figure 4. MAE calculated on the validation dataset by ANFIS
(without updating) model for 1 to 5 day forecasts for:
Verification of output updating procedure (a) entire wet seasons, (b) high water levels.
The comparison of MAE for different forecast lead times

is shown in Figure 3(a) and Figure 4(a) for the entire The incremental updating procedure proposed in the present
training and validation data sets, and for high water levels study significantly improved the accuracy of multi-step-
(> 11 m) in Figures 3(b) and 4(b). The ANFIS models ahead forecasts as shown in Figure 5 and by the time series
show good overall performance as shown in Figures 3(a) plot in Figure 6. The MAE computed on the entire the
and 4(a), since the MAE of the ANFIS model results are wet seasons (Figure 5a) shows only a slight improvement.
significantly less than that for the benchmark and Naïve However, for high water levels (Figure 5b), the improvement
models for all lead times. However, if only high water is more significant, especially for 5-days forecast. The time
levels are considered, MAE is less than the benchmark at series in Figure 6 shows that although there is still a lag-
only 1- and 2- lead-day forecasts, approximately equal to time error, the amplitude error of ANFIS output updated by
the benchmark at 3- and 4-days forecasts, and higher than I-RAR procedure (ANFIS+I-RAR) is lesser, in comparison
the benchmark at 5-days forecast. In the model validation to the original ANFIS model output.
phase, the ANFIS model gives MAE similar to the Naïve 1.25
model at 4- and 5-days forecast for high water levels. Benchmark
1.00 Naïve
1.25 ANFIS
ANFIS+ I-RAR
Benchmark 0.75
MAE (m)
1.00 Naïve
ANFIS
MAE (m)
0.75 0.50
0.50 0.25
0.25
0.00
0.00 (a) 1 2 3 4 5
Lead-time (day)
1 2 3 4 5
(a)
Lead-time (day)
1.25
1.25 Benchmark
1.00 Naïve
Benchmark
1.00 Naïve
ANFIS
MAEhigh (m)
MAEhigh (m)
0.75 ANFIS+ I-RAR

ANFIS
0.75
0.50
0.50
0.25 0.25
0.00 0.00
1 2 3 4 5 (b) 1 2 3 4 5
(b) Lead-time (day) Lead-time (day)
Figure 3. MAE calculated on the training dataset by ANFIS Figure 5. MAE calculated for 1 to 5 lead time steps on
(without updating) model for 1 to 5 day forecasts for: validation data with output updating: (a) entire wet seasons,
(a) entire wet seasons, (b) high water levels. (b) high water levels
48
CONCLUSIONS
The following can be concluded from this study:

1. The ANFIS model is able to provide accurate forecasts
within the benchmark level, for up to 5 -days forecast
for the entire flood season. For the high water forecasts,
ANFIS is accurate only up to 4 lead-days.
2. Significant improvements to 4- and 5-day forecasts
Figure 6. Measured and computed (ANFIS and ANFIS+I-RAR) were obtained when ANFIS model results were updated
time series of 5-lead day forecasts (at the high river stages with the I-RAR output update algorithm.
in 2000 and 2001).
3. Comparing between ANFIS+I-RAR and URBS model
results for 2009, the ANFIS+I-RAR model produced
Comparison of ANFIS and URBS models significantly better results compared to the URBS
model for 1- to 3-days forecasts. However, ANFIS+I-
ANFIS+I-RAR and URBS model results are compared for RAR achieved only slight improvement in comparison
the 2009 wet season in Figure 7. The Naïve model is used with URBS model for 4- and 5-days forecast.
as a baseline for comparison. All error indexes show that the
ANFIS+I-RAR model have significantly better performance
than that of the URBS model at 1- to 3-lead-day forecasts. ACKNOWLEDGMENTS
However, ANFIS+I-RAR achieves only slight improvement
in comparison with URBS model at 4- and 5-days forecast. We wish to thank the Regional Flood Management and
The ANFIS model is able to associate the water levels Mitigation Center - Mekong River Commission for
at Savanakhet and Pakse well for a few days due to the providing the measured data and results from the URBS
high correlation of water levels at Savanakhet and Pakse model.
from 1 to 3 lags. This also implies that the performance
of the ANFIS model will deteriorate when this correlation
decreases as is the case when L > 3 days. It is expected REFERENCES
that ANFIS model results can be further improved by the
inclusion of rainfall as an additional input, to improve 4- [1] A pirumanekul, C., 2006. “Flood forecasting in the Mekong
and 5-days forecasts. This is left as future work. River Basin: an improvement plan for the flood forecasting
system”. International Conference on Mekong Research for
Successful Forecast Percentage
100 100 the people of the Mekong, 18-21 October 2006, Chiang Rai,
80 80 Thailand.
COE (%)
60 60 [2] Box, G.E.P. and Jenkins, G.M., 1976. “Time Series Analysis:
(%)
40 40 Forecasting and Control”. Holden-Day: Oakland, CA.

20 20
[3] Chen, S.H., Lin, Y.H., Chang, L.C. and Chang, F.J., 2006.
0 0 “The strategy of building a flood forecast model by neuro-
1 2 3 4 5 1 2 3 4 5
fuzzy network”. Hydrological Processes, 20: 1525-1540.
Number of Lead Days Number of Lead Days
[4] Goswami, M., O’Connor, K.M., Bhattarai, K.P. and Shamsedin,
A.Y., 2005. “Assessing the performance of eight real-time
1.0 1.2 updating models and procedures for the Brosna River”.
Hydrology & Earth System Sciences, 9(4), 394-411.
RMSE (m)
0.8
MAE (m)
0.5
[5] Jang, J.S.R., 1993. “ANFIS: adaptive-network-based fuzzy
0.4 inference system”. IEEE Transactions on Systems, Man and
Cybernetics, 23(3), 665-685.
0.0 0.0 [6] Ljung, L., 1999. “System Identification: Theory for the User”.
1 2 3 4 5 1 2 3 4 5
Prentice Hall PTR, Upper Saddle River, NJ.
Number of Lead Days Number of Lead Days
[7] Mekong River Commission (MRC), 2005. “Overview of the
15
Hydrology of the Mekong Basin”. Mekong River Commission,
Vientiane, 73 p.
MPAE (%)
10
[8] Mekong River Commission (MRC), 2009. “System
5 Performance Evaluation Report, The MRC Technical Task
Group for verification of the new MRC Mekong Flood
0 Forecasting System (FEWS-URBS-ISIS)”. Mekong River
1 2 3 4 5
Commission Regional Flood Management and Mitigation
Number of Lead Days
Centre, Phnom Penh, Cambodia, October 2009.
Figure 7. Comparison of performance between URBS,
ANFIS+I-RAR and Naïve models [9] MATLAB, 2008. “User guide for release R2008b”. The
MathWorks, Inc.
49
[10] Nayak, P.C., Sudheer, K.P., Rangan, D.M. and Ramasastri, [12] Takagi, T. and Sugeno, M., 1985. “Fuzzy identification of
K.S., 2005. “Short-term flood forecasting with a neuro-fuzzy systems and its applications to modeling and control”. IEEE
model”. Water Resources Research, 41: 1-16 Transactions on Systems, Man and Cybernetics, 15(1), 116-
[11] Serban, P. and Askew, A.J., 1991. “Hydrological forecasting 132.
and updating procedures”. IAHS (International Association [13] WMO, 1992. “Simulated Real-time Inter-comparison of
of Hydrological Sciences) Publication, Vienna, Austria. Hydrological Models”. World Meteorological Organization,
Operational Hydrology Report No. 38. WMO-No. 779.
Geneva Switzerland.
50
DESIGN OF BRINE OUTFALL FOR

SEA WATER REVERSE OSMOSIS (SWRO)
DESALINATION PLANTS
Adrian Law Wing-Keung (cwklaw@ntu.edu.sg)
Shao Dongdong (SHAO0004@ntu.edu.sg)
ABSTRACT: The design of brine outfall for SWRO desalination plants is a complex engineering task. In this paper, three different
design aspects which the authors have conducted recent research are discussed. They are (a) minimizing the recirculation between the
sea water intake and brine outfall, (b) predicting the mixing of brine discharges in coastal waters, and (c) avoiding the Coanda effect at
the discharge point with sufficient clearance from the bottom.
INTRODUCTION The build-up of ambient salinity near the intake can be

attributed to two effects that occur simultaneously, namely
Sea water desalination with Reverse Osmosis (SWRO) near-field and far-field. Near-field recirculation refers to
is an increasingly viable option to supplement the water the active mixing whereby the brine plume spreads and
supply for many coastal cities. Significant advancements propagates in the ambient coastal waters due to the initial
have been made in recent years in membrane technology, momentum and buoyancy fluxes, in such a manner that
process control and brine management, and together they leads to a direct and immediate re-entrainment of portion
have led to a considerable reduction in the water production of the brine by the intake. Adverse near-field recirculation
cost. The output capacity of SWRO plants is also increasing can normally be avoided by properly selecting the outfall
rapidly to achieve an economy of scale, and the brine flow geometry, so that sufficient near-field dilution can occur
rate to the sea has increased correspondingly. This brings in the immediate vicinity of the outfall and the ambient
to focus the need to design properly the brine outfall from standard can be met within a prescribed mixing zone. A
the SWRO plant so that the impact to the environment good description on the near-field processes can be found
can be minimized. in CORMIX (Doneker and Girka, 2001).
The design of brine outfall for SWRO desalination plants Far-field recirculation refers to the residual salinity increase
is a complex engineering task. It involves the necessary in the ambient waters due to the long term buildup of
considerations to determine the discharge location (and salinity by the desalting operation. The magnitude of far-
thus the length of the outfall pipe), the geometrical layout field recirculation depends on the various local ambient
and arrangement (incorporating the bottom bathymetry and characteristics, including the amount of tidal flushing in the
the characteristics of tidal hydrodynamics in the coastal area, the seabed bathymetry and the shoreline geometry.
waters), the type of outfalls (single ports, or multi-ports Shao et al. (2008) and Shao and Law (2009) presented
diffuser), and the range of design flow rates. In what follows, analysis to illustrate that the distance between the outfall and
three aspects that we have conducted recent research are intake is the key factor controlling the amount of far-field
described. recirculation for a uniform water depth. With a complex
bathymetry, however, the situation is more complicated and
the residual salinity can be accumulated in pockets of low
lying seabed undulations. A good far-field simulation using
MINIMIZING RECIRCULATION BETWEEN a comprehensive numerical model would then be essential
SEA WATER INTAKE AND BRINE OUTFALL for analysis (Bleninger et al., 2010).
The brine discharge from the outfall can potentially lead

to an increase in the salinity of the ambient water at the PREDICTING THE MIXING OF BRINE
sea water intake location due to recirculation. Law (2011)
DISCHARGES
performed a quantitative analysis to show that the effect
of recirculation to desalination plants is of high concern,
The near-field mixing of the brine discharges is the major
since the increase in ambient salinity near the intake has
design consideration as far as the structure of the outfall
a direct impact on the operating cost of the desalting
is concerned. Figure 1 shows that the typically near-field
process. Thus, it is important to minimise the recirculation
mixing of a single port brine discharge. The ideal would
for desalination plants, and the potentially higher initial
be for the brine plume to mix effectively with the ambient
capital cost required can be compensated by corresponding
water, to have a maximum rise below the water surface,
savings in the operation cost in the long run.
51
and later to impact the seabed at sloping bathymetries that z0/LM < 0.2, where z0 is the centre height of the port, and
direct the heavier plume towards the open sea. LM is a discharge momentum length scale. For 45°, the
boundary effect was considerably weaker and did not seem
Integral analysis has been widely used to predict the near to affect significantly even for z0/LM down to the smallest
field mixing. In this approach, the governing equations are tested value of 0.05.
first established in terms of continuity, Navier-Stokes and
scalar advection-diffusion equations. To solve the system of
equations numerically, another closure equation is required CONCLUSIONS
and the entrainment hypothesis is often adopted. In the
literature, distinctly different values of the entrainment This bulletin article describes the recent research conducted
coefficient had been reported for buoyant plumes that are by the authors towards the design of brine outfalls for
along or against the direction of gravity. This poses difficulty SWRO desalination plants. Further studies are ongoing to
to the analysis as the brine plume would experience both address other design issues that may affect the operational
regimes through the rising and then falling stage. Shao et performance of the outfall. (The first author is currently
al. (2010) developed a new integral model with a unified a member of the Joint IAHR/IWA Committee on Marine
entrainment function for arbitrary buoyancy that is able Outfalls as well as the Task Group Leader on Brine
to resolve the entrainment non-uniformity in a continuous Outfalls).
manner. Based on extensive comparison with available
data, the new model is shown to be able to be able to
reasonably predict the behavior for the brine discharges REFERENCES
at various inclinations.
[1] Bleninger, T., Niepelt, A., Jirka, G.H., Lattemann, S., Purnama
A., Al-Barwani, H.H. and Doneker, R.L., 2010. “Environmental
AVOIDING COANDA EFFECT AT THE hydraulics framework of the design of discharges from
DISCHARGE PORT desalination plants”. Proceedings of the 6th Int. Sym. Env.
Hyd., Athens, Greece.
In the literature, a 60° inclined single port outfall had [2] Doneker, R.L. and Jirka, G.H., 2001. “CORMIX-GI systems
been recommended for desalination plants to achieve a for mixing zone analysis of brine and wastewater disposal”.
Desalination, 139: 263-274.
maximum mixing efficiency. However, the terminal rise
associated with 60° is relatively high, and thus a smaller [3] Law, A.W.K., 2011. “Recirculation between intakes and
inclination is often more desirable for shallow coastal outfalls of desalination plants”. To presents in the Qingdao
International Desalination Conference.
waters. At the same time, with the smaller inclinations, a
port that is placed at close proximity to the bottom may [4] Shao, D.D. and Law, A.W.K., 2009. “Salinity build-up due to
brine discharges into shallow coastal waters”. Modern Physics
suffer from the Coanda effect that reduces the amount of
Letters B, 23(3): 541-544.
mixing compared to an unbounded environment.
[5] Shao, D.D. and Law, A.W.K., 2010. “Mixing and Boundary
interactions of 30 and 40 degree inclined dense jets”. Journal
Shao and Law (2010) investigated experimentally the of Environmental Fluid Mechanics, 10(5): 521-553.
mixing behavior of brine discharges at smaller angles of
[6] Shao, D.D., Law, A.W.K. and Adams, E.E., 2010. “Integral
30° and 45° in a stationary ambient. Based on the results, modelling of inclined round turbulent jets with arbitrary
the characteristic geometrical features of the inclined dense buoyancy”. Journal of Fluid Mechanics, currently under
jets are quantified. The mixing and diluting behaviors are review.
also revealed through the analysis of the velocity and [7] Shao, D.D., Law, A.W.K. and Li, H.Y., 2008. “Brine discharges
concentration profiles. The study also examined the effect into shallow coastal waters with mean and oscillatory tidal
of proximity of the discharge port to the bed. For 30°, it currents”. Journal of Hydro-environment Research, 2(2): 91-
was found that the bed influence became significant when 97.
x/D
Figure 1. Typical pattern of the mixing of a brine discharge from an inclined port.
52
FOULING BEHAVIOR OF FORWARD

OSMOSIS MEMBRANES
She Qianhong (QHSHE@ntu.edu.sg)
Gu Yangshuo (YGU4@e.ntu.edu.sg)
Tang Chuyang (CYTang@ntu.edu.sg)
ABSTRACT: Forward Osmosis (FO), an emerging separation technology, has potential applications in water and wastewater treatment as
well as desalination. Internal concentration polarization (ICP) and membrane fouling have been found to adversely affect its performance.
This research aims to study the fouling behavior of FO membranes using humic acid as a model foulant. By performing batch cross flow
fouling experiments, the effect of FO membrane orientation and draw solution concentration on the fouling behavior were investigated.
Compared to the active layer facing draw solution (AL-facing-DS) orientation, flux was remarkable stable for the active layer facing
feed solution (AL-facing-FS) orientation due to the ICP self-compensation effect. FO suffered greater flux loss in AL-facing-DS, which
was likely due to the internal clogging of the porous structure and the resulting enhanced ICP in the support layer.
INTRODUCTION This study aims to systematically investigate fouling

behavior in the FO process and further explore the potential
Forward osmosis (FO), which uses osmotic pressure mechanisms inherent in the FO process.
difference across a selective membrane as the driven force,
has gained growing interests in both research and application
in recent years (Cath, Childress et al. 2006; Tang, She et al. MATERIALS AND METHODS
2010; Wang, Shi et al. 2010; Xu, Peng et al. 2010). Due
to its prominent advantages of low energy consumption In the current research, the FO membrane was derived
and high solute rejection, FO has been applied in the from the commercial Hydrowell filter which was purchased
water and wastewater treatment and seawater or brackish from Hydration Technology Inc. (HTI). The purified
water desalination (Cath, Childress et al. 2006). In the Aldrich humic acid (PAHA) was used as a model foulant.
FO process, internal concentration polarization (ICP) has The schematic diagram of the FO membrane filtration
been identified to be a major obstacle of the FO membrane bench-scale experimental system is shown in Figure 1.
performance in previous studies (Mehta and Loeb 1978; The system was comprised of two loops of solutions (i.e.,
Lee, Baker et al. 1981; McCutcheon and Elimelech 2006; feed solution and draw solution) which were separated by
Tang, She et al. 2010). The water flux Jv in a different the FO membrane in the cross flow membrane filtration
membrane orientation due to the ICP effect can be modeled cell. The water flux of the FO process was gained by
by the Eq. (1) and Eq. (2) (Tang, She et al. 2010). measuring the weight changes of the feed solution in the
fixed duration. Unless otherwise specified, the following
Jv = Kmln ( AπAπ + J+ +B B) (AL-facing-FS)
draw
feed v
(1) conditions were used as the reference conditions: 10 mM
NaCl and 10 mg/L humic acid in FS, crossflow velocity
23.2 cm/s, temperature 22-24oC.
Jv = Kmln (AπAπ – J+ B+ B) (AL-facing-DS)
draw
feed
v
(2)
where A and B are the respective transport coefficients for

water and solute, πfeed and πdraw are the respective osmotic
pressure of the feed solution and draw solution, Km is the
solute mass transfer coefficient, which is given by
D D
Km = l = (3)
τ. ε S
Where D is the solute diffusion coefficient, τ is the tortuosity
of the support layer, l is the length of the support layer, ε
is the porosity of the support layer, and S is the membrane
structure parameter.
Fouling is another important drawback limiting the flux

level of the FO process (Mi and Elimelech 2008; Lay,
Chong et al. 2010; Tang, She et al. 2010). However, few Figure 1. Schematic diagram of the FO membrane
publications have systematically studied the mechanisms of filtration system.
FO membrane fouling so far (Mi and Elimelech 2008).
53
50 AL f acing 40
45 draw solution
AL facing FS
40 AL f acing Fouling test Baseline
35
35
f eed solution AL facing DS
Flux(g/10 min)
30
Fouling test Baseline
30
25
20
25
Flux (L/m hr)

15
2
10
5 20
0
0 200 400
00 600 800 15
ti ( iin))6
tim e(m
Figure 2. Flux behavior in the baseline flux tests. 10
5
During the whole test, the DS was diluted and the FS was 0 60 120 180 240 300 360 420 480
concentrated as the water was transported from the FS to
Time (min)
the DS. Therefore, prior to each fouling experiment, the
baseline test which has an identical condition except without Figure 3. Effect of the membrane orientation on the fouling
behavior of FO membrane (cited from reference
adding foulant in the FS was performed to identify the
(Tang, She et al. 2010)).
flux loss due to membrane fouling. Figure 2 presents a
group of baseline fluxes under the two types of membrane
orientation using 2 M NaCl as draw solution. Due to the conducted in the AL facing DS configuration. Clearly,
severe ICP in the AL-facing-FS, the flux in the AL-facing- greater baseline flux level was obtained at higher draw
FS was much lower than that in the AL-facing-DS. solution concentration as a result of the increased osmotic
pressure difference at higher draw solution concentration.
For the fouling tests, the flux level exhibited nearly identical
RESULTS AND DISCUSSION trend with the baseline flux at 0.5 M and 1 M draw solution
concentration (low initial flux level), while significant flux
Effect of FO membrane orientation loss was observed when the draw solution concentration
was 2 M and 4 M (initial flux was above 40 L/m2h).
Figure 3 illustrates the FO fouling behavior under the two The phenomenon of severe fouling behavior at high draw
different membrane orientations. The experiments at the solution concentration (thus high driven force) in the FO
two different membrane orientations were conducted under process was similar to that in the pressure-driven membrane
the same initial flux (~30L/m2.hr). As shown in Figure process. This is attributed to the greater hydrodynamic
3, little flux reduction was observed for the AL-facing-FS drag force for promoting foulant deposition.
configuration, while the flux decreased significantly for the
AL-facing-DS configuration. When the membrane rejection 60
Baseline, 4 M Fouling, 4 M
Baseline, 2 M Fouling, 2 M
layer faced the feed solution, humic acid had lower tendency Baseline, 1 M Fouling, 1 M
55
to deposit on the smooth surface, which was verified by Baseline, 0.5 M Fouling, 0.5 M
the minimal measured humic acid deposition. As a result, 50
no significant flux reduction occurred. In addition, the ICP 45
self-compensation effect is another plausible explanation for 40

the flux stability under the AL-facing-FS configuration.
Flux (L/m hr)
35
2
30
For the AL-facing-DS configuration, the porous support
25
layer was exposed to the feed solution, humic acid had
greater tendency to enter the porous support layer and was 20
retained inside the support by the active layer, inducing 15

severe internal pore clogging. This effect significantly 10
increased the structure parameter (S value) and thus reduced 5
mass transfer coefficient Km. As the ICP is exponentially 0 60 120 180 240 300 360 420 480
dependent on the mass transfer coefficient, the observed Time (min)
flux decreased significantly.
Figure 4. Effect of the draw solution concentration on the
fouling behavior of FO membrane (cited from reference
Effect of the draw solution concentration (Tang, She et al. 2010)).
The effect of draw solution on FO fouling behavior is

shown in Figure 4. The baseline and fouling tests were
54
CONCLUSIONS [4] McCutcheon, J.R. and M. Elimelech (2006). “Influence of

concentrative and dilutive internal concentration polarization
The fouling behavior in the FO process was systematically on flux behavior in forward osmosis.” Journal of Membrane
Science, 284(1-2): 237-247.
investigated. In the AL-facing-FS orientation, the flux
exibted inherently stability. Any attempt to decrease the [5] Mehta, G.D. and S. Loeb (1978). “Internal polarization in
flux level was compensated by the reduced level of ICP. the porous substructure of a semipermeable membrane under
pressure retarded osmosis.” Journal of Membrane Science,
In contrast, in the orientation of AL-facing-DS, the flux
4(2): 261-265.
was subject to large loss under the fouling conditions due
[6] Mi, B. and M. Elimelech (2008). “Chemical and physical
to the pore clogging enhanced ICP as well as the reduced
aspects of organic fouling of forward osmosis membranes.”
membrane permeability. The enhanced ICP effect was Journal of Membrane Science, 320(1-2): 292-302.
more dominant at higher flux levels.
[7] Tang, C.Y., Q. She, et al. (2010). “Coupled effects of internal
concentration polarization and fouling on flux behavior of
forward osmosis membranes during humic acid filtration.”
REFERENCES Journal of Membrane Science, 354(1-2): 123-133.
[8] Wang, R., L. Shi, et al. (2010). “Characterization of novel
[1] Cath, T.Y., A.E. Childress, et al. (2006). "Forward osmosis: forward osmosis hollow fiber membranes.” Journal of
Principles, applications, and recent developments." Journal Membrane Science, 355(1-2): 158-167.
of Membrane Science, 281(1-2): 70-87.
[9] Xu, Y., X. Peng, et al. (2010). “Effect of draw solution
[2] Lay, W.C.L., T.H. Chong, et al. (2010). “Fouling propensity concentration and operating conditions on forward osmosis
of forward osmosis: Investigation of the slower flux decline and pressure retarded osmosis performance in a spiral wound
phenomenon.” Water Science and Technology, 61: 927-936. module.” Journal of Membrane Science, 348(1-2): 298-309.
[3] Lee, K.L., R.W. Baker, et al. (1981). “Membranes for
power generation by pressure-retarded osmosis.” Journal of
Membrane Science, 8(2): 141-171.
55
ESTIMATE OF RESISTANCE INDUCED

BY SIMULATED EMERGENT VEGETATION
IN OPEN CHANNEL FLOWS
Nian-Sheng Cheng (cnscheng@ntu.edu.sg)
Hoai Thanh Nguyen (c070041@ntu.edu.sg)
INSTRUCTION Table 1 compares rv with other hydraulic radii. Using rv, the
Colebrook-type resistance relation for open channel flows
The vegetation drag and its relevant Reynolds number have subject to emergent vegetation is proposed here,
been defined diversely in the literature and some definitions
are even misleading because of improper use of length fv = f (Rev) …(3)
and velocity scales. As a result, a general formula similar 8grvS 4V r
to Manning equation developed for regular open channel where fv = 2 and Rev = νv v .
Vv
flows is not available at present for evaluating resistance
to vegetated open channel flows, even for the emergent
case that is relatively simple.
In this study, a vegetation-related hydraulic radius is

proposed and applied to redefine the Reynolds number
and friction factor for vegetated open channel flows. All
considerations are limited to the case of emergent rigid
vegetation simulated with circular cylinders.
VEGETATION-RELATED HYDRAULIC
RADIUS
(a) Plan view (b) Side view
As shown in Fig. 1, we consider emergent vegetation Figure 1. Emergent vegetation simulated with
simulated with staggered rigid cylinders. In this case, the circular cylindrical rods.
effective vegetation height is the same as the flow depth h,
and the configuration of vegetation is solely governed by Table 1. Length scales for characterizing flow geometry.
the stem diameter d and density λ defined as the average
volume fraction occupied by stems. Geometrical
Flow geometry Hydraulic radius
dimension
For vegetated flows without sidewall and bed effects, we D
Circular pipe pipe diameter, D
4
consider a 3D domain that measures ∆x × ∆y × 1, in the -1
Rectangular channel width, B 1 1
streamwise, lateral and vertical direction, respectively. +
open channel flow depth, h h 0.5B
Similar to the regular hydraulic radius, the hydraulic radius
Porous media grain size, d50 11–λ
in the 3D domain is defined as the ratio of the volume d50
comprising grains pore size 6 λ
occupied by water to the wetted surface area,
Vegetated channel
stem diameter, d π 1 – λ
without channel d
…(1) stem spacing, s 4 λ
boundary effects
channel width, B
-1
Note that in the above definition, the total wetted surface Open channel § 1 ·
flow depth, h ¨1 1
area is used. For vegetation-induced form drag, we only with emergent ¨h + + π 1 – λ ¸¸
stem diameter, d ¨ 0.5B d¸
need to consider the frontal area of the stem. This yields vegetation © 4 λ ¹
stem spacing, s
an effective wetted area equal to Nd in the 3D domain, and
then the vegetation-related hydraulic radius is given by
…(2)
56
DRAG COEFFICIENT i.e.

2/3
S1/2 Bh
For each cylindrical stem, the drag force in the streamwise n= …(11)
V B + 2h
direction is
Obviously, the n-values so obtained would generally depend
Vv2
FD = CDvρhd …(4) on vegetation configuration for vegetated open channel
2
flows. From Eq. (7),
where CDv is the drag coefficient. The total drag per unit
bed area is
2 2
4λ Vv 2λρhVv
NFD = C ρhd = CDv …(5)
πd2 Dv 2 πd
which is equivalent to the streamwise component of the
gravitational force,
2λρhVv2
CDv = (1 – λ)ρghS …(6)
πd CDv
Here, the shear forces induced by sidewalls and bed

are considered negligible. Otherwise, sidewall and bed
corrections are applied. From Eq. (6),
1–λ gr S
CDv = gSπd = 2 v2 …(7)
2λV2v Vv
Furthermore, by comparing Eq. (7) with the definition of
fv, one gets
Rev
1
CDv = f …(8) Figure. 2. Variation of CDv with Rev.
4 v
DEPENDENCE OF CDV ON REV From Eq. (7),

√S CDv
The analysis was performed with the data collected in this
study and also those available in several other studies. Fig.
V
=
√
2grvε2
…(12)
2 plots all data in the form of CDv against Rev, from which Substituting Eq. (12) into (11), we get
the following observations could be made. First, though CDv Bh 2/3
scattered to some extent, all data points generally follow

the same monotonic decreasing trend of CDv with increasing
n=
√2grv(1 – λ)2 B + 2h
…(13)
Rev. Second, for Rev > 7000, CDv fluctuates within a limited Fig. 3 shows that the Manning coefficients predicted using
range, and could be taken to be constant, i.e. CDv ≈ 1.3. Eq. (13), where CDv is estimated using Eq. (9), agree
Third, it seems that CDv dips in the range of Rev = 7 × reasonably with the measurements, i.e. those determined
103 – 6 × 104. This could be related to possible difference using Eq. (11) with the data provided by Ishikawa et al.
that may exist between the channel configured with the [3], James et al. [4], and those collected in this study.
randomly distributed stems (only related to Tanino and
Nepf’s data[1]) and that with staggered stems. Otherwise,
it could be explained by vortex shedding and surface waves CONCLUSIONS
that take place in vegetated channel flows [2].
This study demonstrates that the concept of hydraulic radius

To empirically describe the relationship of CDv and Rev, a is useful to unify experimental data of resistance to vegetated
best-fit function is proposed here open channel flows, which have been collected for various
bed and vegetation configurations. The hydraulic radius
Rev
CDv = 500.43 + 0.7 1 – exp – …(9) is redefined by taking into account effects of vegetation
Rev 15000
size and density, and channel geometry. It serves as a good
length scale in the definition of drag coefficient, friction
factor and Reynolds number for open channel flows subject
MANNING COEFFICIENT to emergent vegetation.
The Manning coefficient can be determined experimentally

from the bulk flow velocity V [= Q/(Bh)], regular open
channel hydraulic radius Bh/(B+2h), and energy slope S,
57
REFERENCES
[1] James, C.S., Birkhead, A.L., Jordanova, A.A. and O’Sullivan,

J.J., 2004. “Flow resistance of emergent vegetation”. Journal
of Hydraulic Research, Journal of Forest Research, 2004;
42(4): 390-8.
[2] Ishikawa, Y., Mizuhara, K. and Ashida, S., 2000. “Effect of
density of trees on drag exerted on trees in river channels”.
Journal of Forest Research, 2000; 5(4): 271-9.
[3] Tanino, Y. and Nepf, H.M., 2008. “Laboratory investigation
of mean drag in a random array of rigid, emergent cylinders”.
Journal of Hydraulic Engineering-ASCE. 2008 Jan; 134(1):
34-41.
[4] Zima, L. and Ackermann, N.L., 2002. “Wave generation in
open channels by vortex shedding from channel obstructions”.
Journal of Hydraulic Engineering-ASCE. 2002 Jun; 128(6):
596-603.
Figure. 3. Comparison of predicted and measured

Manning coefficients.
58
LIFE CYCLE ANALYSIS OF

OFFSHORE GANGWAY
Nguyen Anh Tuan (atnguyen@ntu.edu.sg)
Cao Thi Viet Phuong (CVPhuong@ntu.edu.sg)
Wang Xikun (CXKWang@ntu.edu.sg)
Dai Ying (DaiYing@ntu.edu.sg)
Gho Wie Min (CWMGho@ntu.edu.sg)
Tan Soon Keat (CTANSK@ntu.edu.sg)
ABSTRACT: The life cycle analysis of an offshore gangway as a means of embarkation and disembarkation for personnel access between
the two vessels or between a work boat and an offshore platform was carried out. The strength integrity of the gangway was first reviewed
to validate its geometrical configuration from the requirement of Health and Safety Executive (2002). Key factors that significant influence
to the environmental impact, including production, processing, operation and activities of the gangway are highlighted.
INTRODUCTION its geometrical configuration from the requirement of

Health and Safety Executive (HSE 2002). The key factors
Gangway is widely used in the offshore and marine industry that contribute to the greenhouse effects as well as the
as a specialized equipment to facilitate safe transportation environmental impact with respect to material production,
of goods and personnel between offshore supply vessels and processing and activities related to the fabrication and
platform/vessel. It is a type of bridge structures specially operation of the gangway are highlighted. A scenario is
designed with heave compensation system at supports to hence developed by integrating these factors to emulate the
minimize the effect due to sea motion. Gangway can be actual condition, which would be useful for manufacturers
constructed from steel or aluminium or a combination of and users to evaluate the life cycle of a gangway and its
both to offer a lightweight structure while not comprising operation in the marine environment.
its high strength and stiffness. It can be designed with
sufficient width to accomodate one- or two-way traffic
movements to suit onsite requirements. Objective of study
Life cycle analysis (LCA) is a method for analysing and The objective of the assessment is to determine the
assessing the environmental impact of a material, product key factors suitable for the LCA of steel gangway and
or service throughout its entire life cycle, usually from associated environmental impact. Parameters considered
the acquisition of raw materials to final decommissioning include carbon (CO2), sulfur oxide (SOx) and nitrogen
and disposal. For the LCA of the gangway, its production, oxides (NOx) emission to air and water and the amount
manufacturing process, transportation, installation, of energy consumed.
operation, application, disposal and recycle after use are
briefly discussed. The environmental factors that contribute
to the greenhouse effect include the effect of emissions MODEL SOLUTIONS METHODOLOGY
to the atmosphere and water as well as the amount of
energy consumed. These factors can be incorporated in the The model solution methodology for the gangway is based
analytical tools based on the life cycle assessment approach on the LCA to assess the potential environmental impacts
to determine the impact of this gangway structure on the associated with the production, processing and activities
environment. throughout its entire life cycle. This method analyses the
impact of a product over a lifetime from the process of
extraction of raw materials, manufacturing through to the
Scopes of study waste disposal of its various components for recycling or end
of its service life. The factors relevant to the entire life cycle
The life cycle evaluation of an offshore steel gangway for of the gangway system are compiled to create a scenario
means of embarkation and disembarkation for personnel to show the extent of impact to the environment.
access between the two vessels or between a work
boat and an offshore platform is carried out. Extensive The LCA is conducted based on the guidelines provided in
reference to the published data and information is drawn. the International Organization for Standardization (ISO14040
The strength integrity of the gangway is first reviewed 1996, ISO14041 1998, ISO14042 2000, ISO14043 2000
with the application of finite element analysis to validate and ISO14044 2006). The procedure includes the data
59
collection and evaluation and the interpretation of results used as part of a reference to further investigate in greater
after analysis. In this evaluation of the life cycle of the details the life cycle of various types of offshore gangway
steel gangway, the approach is to firstly gather data and used in the marine industry.
information through the literature review of previous
publications of steel structures based on LCA methodology.
The process flow chart of the major life cycle of gangway
product is summarized as shown in Figure 1. Generally,
the process flow chart for typical gangway comprises five
(5) main stages which can be used as a framework for
inventory analysis. Each stage will contain unit processes
with elementary inputs and outputs represented as building
blocks for data collection.
Figure 2. Emission to the atmosphere (kg).
Figure 1. Process flow chart with data of gangway system.
LIFE CYCLE ANALYSIS OF GANGWAY
Based on the major process flow stages of the gangway

products as shown in Figure 1, all the unit processes within
each of these stages are considered within the system
boundary, and also include all elementary flows that relate
to the environmental impacts. The inputs are the energy
consumption and the natural resources such as water. The
elementary flow outputs in the system boundary are the
waste removal and the harmful air emissions specifically
Figure 3. Discharge to recipient water (kg).
for those that have significant influence on global warming.
Air emissions that relate to toxicity and acidification are
considered as primary effects.
Data for each stage is collected from two main sources.

The first source can be found in the database established by
the Swiss Centre for Life Cycle Inventories which provides
thousands of process flows across the entire spectrum of
manufacture industries (Ecolnvent, 2010). The second is

the industry reports and publications including material
specific journals, reports and websites.
RESULTS
The results of the LCA calculation based on a typical

gangway available in the market in such aspects as emissions
to atmosphere, water and the energy consumptions per unit
weight of 100kg of the product are presented in Figures
2, 3 and 4, respectively. This set of information can be
Figure 4. Energy consumption (MJ).
60
CONCLUDING REMARKS REFERENCES
In this study, the geometrical configuration and the strength [1] Ecolnvent database 2010. http://www.pre.nl/ecoinvent/, The
integrity of the offshore gangway has been verified against Netherlands.
the requirement of Health and Safety Executive. This [2] HSE 2002. Health and Safety Executive, Decks, stairways,
preliminary study was based on extensive reference to gangways and their associated handrails, Offshore technology
the published data and information and was conducted to report 2001/69, Bomel Ltd, UK.
determine important factors that would have contributed [3] SO 14040 2006. International Organization for Standardization,
significant impact to the environment. The results of the Environmental management, Life cycle assessment, Principles
sulfur oxide, nitrogen oxides and carbon emission to the and framework, Switzerland.
atmosphere, recipient water and the energy consumption [4] ISO 14041 1998. International Organization for Standardization,
from the process of gangway manufacture and production Environmental management, Life cycle assessment, Goal and
scope definition and inventory analysis, Switzerland.
are presented.
[5] ISO 14042 2000. International Organization for Standardization,
Environmental management, Life cycle assessment, Life cycle
impact assessment, Switzerland.
Environmental management, Life cycle assessment, Life cycle
interpretation, Switzerland.
Environmental management, Life cycle assessment,
Requirements and guidelines, Switzerland.
61
DAM-BREAK FLOW SIMULATION WITH

SEDIMENT TRANSPORT
Pu Jaan Hui (jhpu@ntu.edu.sg)
Shao Songdong (s.shao@bradford.ac.uk)
Tan Soon Keat (CTANSK@ntu.edu.sg)
ABSTRACT: In this study, 2D shallow water equations together with a sediment continuity-concentration (SCC) model were applied to
flow with mobile sediment boundaries. The SCC model could reproduce the information for both bed and suspended loads. A combination
of the shallow water and the sediment transport models was constructed using the fully conservative law to preserve the integrity of
their governing equations, and the proposed computation model was based on the Finite Volume (FV) method. The Monotone Upwind
Scheme of Conservative Laws (MUSCL)-Hancock scheme was used with the Harten Lax van Leer (HLL) approximate Riemann solver
to discretize the FV model.
INTRODUCTION show the two-dimensional fully conservative shallow water

and sediment continuity-concentration equations [Valiani
A variety of numerical schemes have been used and and Caleffi (2001), Cao et al. (2004), and Wu and Wang
proposed to simulate the sediment laden flows. The sediment (2008)]
continuity model is one of the most basic sediment transport
models to simulate sediment transport in flow. However, …(1)
since this model only considers the bed load sediment
transport, it could not be used to accurately represent
sediment transport flow with rapid suspended concentration …(2)
change, such as debris or dam-break flow.
Realising the shortcomings of the earlier sediment continuity

model, Armanini and Di Silvio (1988) suggested a set
of sediment continuity-concentration (SCC) equations to
improve the sediment transport representation by including …(3)
the exchange effect of the bed and suspended loads. Their
model was solved in a 1D domain but with consideration
of the non-equilibrium lag of sediment transport. These
equations were further tested by Valiani and Caleffi (2001),
and Wu and Wang (2008) for the case of turbulent dam-
break sediment transport flow, and good correspondence …(4)
between the numerical simulation and the experimental
data were observed.
…(5)
Most of the previous sediment laden flow models were
constructed based on the sediment bed load assumption The variable φ refers to geopotential, and is given by
using a 1D flow continuity-momentum approach. In this φ = g • h ; where h is the water depth; g is the gravitational
paper, the combination of 2D shallow water and SCC acceleration. u and v are the depth averaged flow velocities
models were used to simulate the bed and suspended in streamwise and lateral directions respectively; ρs and ρw
loads movement in a 2D depth averaged flow. The are density of sediment and water respectively; C is the
proposed combined model was used to predict a highly flux-averaged volumetric sediment concentration; λ is the
discontinuous dam-break sediment transport flow. The sediment bed porosity; and zb is the bed elevation. es and ds
are sediment erosion and deposition rates respectively. x, y
experimental measurements from the literature were also
and t denote the spatial-longitudinal, spatial-transverse and
used for validation. time domains respectively. Sox and Soy in equations (2) – (3)
are the bed slopes in the streamwise and lateral-directions
respectively; and the friction slope of the channel, Sf, are
GOVERNING EQUATIONS given by
The proposed model described in this paper was built using , and …(6)
the sediment continuity-concentration model combined
with the shallow water flow equations. Equations (1) - (5) where, n in the Manning friction coefficient.
62
NUMERICAL SCHEME term, a first order derivative was used for its discretization.
No complex discretization method is needed as HLL-type
There are different discretization techniques that can be scheme is having competence to capture source term solution
used to discretize a Shallow Water model, namely Finite accurately (Hu et al., 2006).
Difference (FD), Finite Element (FE) and Finite Volume
(FV) methods. In this study, FV method was chosen for
its robust nature in simulating high dimensional shock- RESULTS AND DISCUSSIONS
capturing flow profiles accurately (Mingham and Causon,
2000). A dam-break flow is a well-known and rapidly varying flow
event with a highly unpredictable nature. The effect of a
In the inviscid flux modelling of the proposed FV model, the dam-break flow on a mobile sediment bed was investigated
Godunov-type Hancock scheme was used with a two-stage by Capart and Young (1998) experimentally using a channel
predictor-corrector temporal discretization. The Godunov- equipped with a rapidly opened sluice-gate. This experiment
type Hancock scheme was coupled with Harten Lax van was simulated using the proposed model and the results
Leer (HLL) approximate Riemann solver for upwinding are reported herein.
volumetric discretization. The slope limiter method was
used in the HLL solver to ensure the spatial discretization A rectangular channel with dimensions of 12.0m length
scheme satisfies flux-limiting property (Mingham and and 0.2m width was used for Capart and Young (1998)
Causon, 2000). MUSCL (Monotone Upwind Scheme experimental test case. The sediment used in the test
for Conservation Laws) scheme was used to update the has a size of 6.1mm, a density of 1048kg/m3, and a fall
variables spatial wise. The MUSCL and Hancock schemes velocity of 0.076m/s. The flow was initially at rest, where
keep the proposed FV model at a second order of accuracy it had an initial water depth of 10cm at the upstream and
in spatial and temporal domains respectively. The overall dry water depth at the downstream with a layer of 6cm
inviscid solution is indicated by sediment at the bed throughout the channel from upstream to
downstream. The sluice gate, which was originally situated
…(7) at 4m location from upstream, was opened instantaneously
at the start of the flow. The results of the water surface
For time stepping, the Courant-Friedrichs-Lewy criterion and bed load elevation profiles around the dam breaking
was followed to ensure the utilised time step does not point are shown in Figure 1. The time-distribution of the
exceed its maximum allowable limit. In the numerical source
Figure 1. Water surface and bed elevation profiles comparison for different models
63
flow profiles are shown for every 0.1s interval over 0.5s REFERENCES
to capture the transient changes in the water flow and bed
change conditions. The proposed model results are plotted [1] Armanini, A., and Di Silvio, G., 1988. “A One-dimensional
against the experimental measurements of Capart and Young model for the transport of sediment mixture in non-equilibrium
(1998). The sediment continuity modelling results from conditions”. Journal of Hydraulic Research, 26(3): pp. 275-
Capart and Young (1998), and the sediment continuity- 292.
concentration modelling results from Wu and Wang (2007) [2] Cao, Z., Pender, G., Wallis, S. and Carling, P., 2004.
are also plotted in the same figure for comparison. “Computational dam-break hydraulics over erodible sediment
bed.” Journal of Hydraulic Engineering, 130(7): pp. 689-
Both crucial dam break flow characteristics of water wave 703.
front and bed scour hole size are well-predicted using
[3] Capart, H. and Young, D.L., 1998. “Formation of a jump by
the proposed model compared to the other computational
the dam-break wave over a granular bed”. Journal of Fluid
models. Numerically, the behaviour of the suspended load Mechanics, 372: pp. 165-187.
in the proposed model is simulated using the sediment
concentration partial differential equations. This improved [4] Hu, K., Mingham, C.G. and Causon, D.M., 2006. “A mesh
feature simulates the suspended sediment information at patching method for finite volume modelling of shallow water
the water wave front, hence giving a consistent prediction flow”. International Journal for Numerical Methods in Fluids,
50: pp. 1381-1404.
to the measured water wave front.
[5] Mingham, C.G. and Causon D.M., 2000. “Calculation of
unsteady bore diffraction using a high resolution finite volume
method”. Journal of Hydraulic Research, 38(1): pp. 49-56.
CONCLUSIONS
[6] Valiani, A. and Caleffi, V., 2001. “Dam break modeling for
sediment laden flows”. Proceedings of the 2001 International
A 2D shallow water Finite Volume (FV) numerical model Symposium on Environmental Hydraulics, Arizona, USA,
was developed to analyse shallow flow with sediment pp.1-6.
transport. The proposed model was applied to a dam-break [7] Wu, W. and Wang, S.S.Y., 2007. “One-dimensional modeling
flow test over movable bed in this paper. The proposed of dam-break flow over movable beds”. Journal of Hydraulic
model simulation was compared to the experimental Engineering, 133(1): pp. 48-58.
measurements in literature, and the comparison shows [8] Wu, W. and Wang, S.S.Y., 2008. “One-dimensional explicit
good corresponding results at its water and bed elevation finite-volume model for sediment transport with transient
predictions. The above-mentioned statements are clearly flows over movable beds”. Journal of Hydraulic Research,
demonstrated by the water wave front and the bed scour 46(1): pp. 87-98.
hole size predictions of the proposed model, which are
not predicted so accurately by the other models proposed
in the literature.
64
NUMERICAL SIMULATION OF
WEDGE WATER ENTRY BASED ON
TWO-PHASE SPH MODEL
GONG Kai (gongkai@ntu.edu.sg)
LIU Hua (hliu@sjtu.edu.cn)
TAN Soon Keat (ctansk@ntu.edu.sg)
ABSTRACT: The hydrodynamic problem of two-dimensional wedge entering water was studied based on two-phase SPH model. A
non-reflection boundary treatment for SPH method was proposed to reduce the size of computational domain. The details of water entry
and enclosing were simulated using multi phase SPH model. Numerical simulations verified our experimental observations.
GENERAL INSTRUCTIONS simulating the multiphase flows may improve the force
prediction.
Since the first studies by von Karman, lots of works have
been carried out in water entry problems. Water entry of
a solid through the free surface is still a current focus of NUMERICAL MODELING
the researches in hydrodynamics. The water entry problem
is part of the general fluid-structure impact problem. The The momentum and kinematics equations for fluid particles
first complete solution was obtained by Dobrovol’skaya in SPH method are
(1969) for a two-dimensional wedge based on velocity
potential theory. Zhao and Faltinsen (1993) studied the …(1)
same problem using a refined procedure with a more
advanced computer.
Smoothed Particle Hydrodynamics (SPH) is a meshfree

method that offers substantial potential in many classes
of problems especially those characterized by large …(2)
deformations. It was originally developed for astrophysical
computations (Gingold and Monaghan 1977, Lucy 1977)
where a represents the reference particle; b represents the
and has later been extended to model a wide range of
neighboring particles of a; Wab = W(|ra – rb|, h) is kernel
problems including multi-phase flow (Monaghan and
function defined by
Kocharyan 1995), impact and fracture problem (Randles et
al. 1995). It is a pure Lagrangian, particle method. Unlike
the Particle In Cell method (PIC), SPH method does not
need a grid to calculate spatial derivatives. Instead, it is
based on analytical differentiation of interpolation function.
The continuum equation and the momentum conservation
equations are reduced to sets of ordinary differential
equations. The particle positions and attributes are computed
using standard numerical integration methods in time
domain such as leap-frog scheme.
The SPH method has extraordinary potential for problem

solving in areas where traditional techniques have been
…(3)
unsuccessful. The method has the advantage of conceptual
simplicity, ease of implementing new physics, natural
For weak compressible method, the pressure was
treatment of void regions, ease of simulating three-
calculated by state equation that was first used by
dimensional problems and ease of handle large deformations
Monaghan (1994) and based on Batchelor’s (1974) equation,
of the free surface. SPH method is attractive on simulating
P = P0 (ρ – ρ0)γ – 1 , where γ = 7 and ρ0 = 1000 for water.
the violent wave impact problems, e.g. Oger et al (2006).
The parameter P0 was chosen to have maximum density
For water entry problems, the air cavity enclosed by the
oscillations of order of O(1%) around the reference density
water may significantly affect the local free surface profile
ρ0. In practice, the constant P0 refers to the sound speed
and flow field, and then the hydrodynamics loads. Correctly
65
c02 = dP / dρ which is ten times or larger than the highest

fluid velocity expected in the physical problem, that is
P0 = ρ0 c02 / γ. For weak compressible flows, the density
was calculated by integrating:
dρa
dt
= – ρa∇ua = ∑ mb (ua – ua)∇aWab …(4)
b
The two-phase SPH model given by Colagrossi and Landrini

(2003) was adopted. Non-absorbing boundary Absorbing boundary
Figure 2. Pressure distribution of the whole flow field.
CODE VERIFICATIONS
SIMULATION OF CAVITY ENCLOSING
Verification was conducted to verify the weak compressible
SPH code and various boundary implementations. A two-
dimensional wedge vertical impacting the free surface is In some circumstances, the single phase model could not
simulated. The boundary pressure was obtained using an predict the physical process after enclosing, wherein the
improved coupling boundary treatment approach. Three pressure is zero even in the enclosed cavity. However, with
snapshots of pressure distribution during water entry are two phase SPH model, the flows of entrapped air could be
illustrated in Figure 1. Computational results were compared well simulated after enclosing of the cavity. A comparison
with the experimental and analytical results given by Zhao between numerical results and physical experiments shows
& Faltinsen (1993). The pressure distribution fits well with a good agreement, as shown in the particles distribution in
the analytical results. Figure 3. Only water particles are shown in the figure.
To save computational efforts, absorbing boundary was
implemented to remove the sound disturbance from the
computational domain, details in Gong et al (2009). By
this approach, the computational time can be extended
without the limitation of sound wave’s reflection from
solid boundaries, which is a common drawback of weak
compressible SPH method. Figure 2 shows the effects of
absorbing boundary. The sound wave’s reflection from solid
boundaries is obviously reduced.
Figure 3. Comparison between experimental results and

computed results of water phase.
The flow field and pressure is shown in Figure 4. The

pressure of vented cavity equals the atmosphere pressure,
shown in relative value. After enclosing, the pressure in the
sealed cavity increases rapidly and equals to the surrounding
water’s pressure. Note the re-entrant jet is formed after the
deep enclosure.
Figure 1. Distribution of pressure for water entry of wedge in

the case of free moving. (top: t = 0.00438s;
middle: t = 0.0158s, bottom: t = 0.0202s).
66
REFERENCES
[1] Batchelor, G.K., 1974. “Introduction to Fluid Dynamics”.

Cam. Univ. Press, Cambridge, U.K.
[2] Colagrossi, A. and Landrini, M., 2003. “Numerical Simulation
of Interfacial Flows by Smoothed Particle Hydrodynamics”.
J. Comput. Phys., 191: 448-475.
[3] Dobrovol’skaya, Z.N., 1969. “On some problems of similarity
flow of fluid with a free surface”. J. Fluid Mech., 36: 805-
829.
Figure 4. Pressure (in colour) and velocity vector [4] Gingold, R.A. and Monaghan, J.J., 1977. “Smoothed particle
distribution before and after cavity closure. hydrodynamics: Theory and application to non-spherical
stars”. Mon. Not. R. Astron. Soc., 181: 375-389.
CONCLUSIONS [5] Gong, K., Liu, H. and Wang, B.L., 2009. “Water Entry of a
Wedge Based on SPH Model with an Improved Boundary
Treatment”. Journal of Hydrodynamics, 21(6): 750-757.
To simulate the enclosing process after water entry, a
[6] Lucy, L.B., 1977. “A numerical approach to the testing of
two-phase SPH method was applied. With fine particle
the fission hypothesis”. Astron. J., 82(12): 1013-1024.
distribution, details of water entry, including surface
[7] Monaghan, J.J., 1994. “Simulating Free Surface Flows with
profile, pressure distribution and total force etc, could
SPH”. J. Comput. Phys., 110(2): 399-406.
be well predicted. The enclosing of water entry was
successfully simulated with proposed SPH model, providing [8] Monaghan, J.J. & Kocharyan, A., 1995. “SPH simulation of
multi-phase flow”. Computer Physics Communication, 87:
a powerful Lagrangian approach for violent free surface
225-235.
flow calculation.
[9] Oger, G., Doring, M., Alessandrini, B. and Ferrant, P., 2006.
“Two-dimensional SPH simulations of wedge water entry”.
J. Comp. Phys., 213: 803.
[10] Randles, P.W., Carney, T.C., Libersky, L.D., Renick, J.R.
and Petschek, A.G., 1995. “Calculation of oblique impact
and fracture of tungsten cubes using smoothed particle
hydrodynamics”. International Journal of Impact Engineering,
17: 661-672.
[11] Zhao, R. and Faltinsen, O., 1993. “Water entry of two-
dimensional bodies”. J. Fluid Mech., 246: 593-612.
67
OPTIMIZATION AND ENHANCEMENT

OF MICROBIAL HYDROLYSIS OF
LIGNOCELLULOSIC WASTE TO
REDUCING SUGARS
Bernard Ng Jia Han (bernhan@ntu.edu.sg)
Wang Jing-Yuan (jywang@ntu.edu.sg)
Qi Wei (qiwe0001@ntu.edu.sg)
Chen Chia Lung (clchen@ntu.edu.sg)
ABSTRACT: Nineteen pure strains of lignocelluloses-converting microorganisms were successfully isolated in this study. The best
combination of pure culture of Microbacterium sp. F28, Tsukamurella sp. C35, Pseudallescheria sp. D42 and Bacillus sp. F4 showed
the maximum reducing sugars yield of 1,653 mg/L, which is ten times higher than the original consortia. The maximum reducing sugars
production yield of 173 mg reducing sugars/g lignocellulose was obtained under the optimum conditions at temperature 56.9 ºC, pH 5.8
and with initial lignocellulose concentration at 28.9 g/L after orgnosolv pretreatment.
INTRODUCTION soil (GS). Samples were taken from the four sources in
sterilized sample bags and stored at 4 ºC until cultivation of
The world’s energy crisis has exacerbated in recent years. microorganisms was performed. The inoculated flasks were
Alternative non-fossil fuel energy sources are in urgent fed with lignocellulose as the sole carbon source with the
demand. Bioethanol is such an alternative fuel which has following basic nutrient: (NH4)2SO4 2 g L-1, CaCl2 0.1 g·L-1,
low carbon content and environmentally friendly. However, KH2PO4 0.5 g·L-1, K2HPO4 2.0 g·L-1, MgSO4·7H2O 0.1 g·L-1,
current food-based feedstock for bioethanol production may NaCl 6.0 g·L-1, Yeast extract 1.0 g·L-1 and lignocellulose 10
not be practical since it may cause competition between g·L-1. Tenebrio Molitor Linnaeus (yellow mealworm) was
food and energy crops for land use. A potential source of provided. Every 8 mealworms were disrupted by passing
low-cost ethanol production is to utilize lignocellulosic through a syringe needle for each medium. The disrupted
waste, viz crop residues, grasses, sawdust, wood chips and gut debris were suspended in 200 ml medium 1 [2]. Medium
solid animal waste, which represents one of the foreseeable 1 and Medium 2 were supplemented with carboxymethyl
sustainable sources in nature because of its relatively low cellulose (CMC; 10 g L-1; 21900; Fluka, Buchs, Switzerland;
cost and plentiful supply [1]. Hydrolysis of lignocellulosic degree of substitution 0.70-0.85; medium 1a), filter paper
waste to reducing sugars is one of the key steps during strips (10 g L-1, brand; medium 1b), and lignocellulose waste
bio ethanol production. The main drawbacks of traditional powder (10 g L-1; with a size of less than 1 mm; medium
hydrolysis methods (viz enzymatic and acidic treatments) 1c), respectively. The flasks were placed in a reciprocal
are their high operation costs, generation of secondary incubation shaker (Grant, OLS 200, Grant Instruments,
pollution and their under optimized and underdeveloped UK) at 30±2°C and at a horizontal rotational speed of 150
technologies. Thus, in this study, microbial hydrolysis of rpm. Samples in the cultivated flasks were analyzed every
lignocellulosic waste to reducing sugars was employed. 1- 24 hours in the cultivated flasks and measurements were
year old compost, 4-month old compost, aerated activated recorded. pH was measured with a pH meter (CyberScan
sludge, garden soil and yellow mealworm guts were selected PCD 6500 pH meter, Germany). A phenol-sulfuric acid
as inoculate sources to isolate lignocellulose-converting method was used to quantify polysaccharides [3], with
microorganisms. Enhancement of microbial hydrolysis of glucose as the standard. Monosaccharides were measured
lignocellulosic waste to reducing sugars was investigated by DNS method with glucose as the standard [4]. Each
by the combination of pure strains and optimization of sample was measured in triplicates.
environmental conditions to further improve reducing
sugars yield.
MATERIALS AND METHODS Nineteen dominant microbes were successfully isolated

and characterized from microbial communities of the
The inoculums originated from four specifically selected four natural sources and mealworm guts. Cellulomonas
natural sources, i.e., 1-year compost (C1), 4-month compost and Pseudomonas were the most abundant phylum in all
(C4), aerated activated sludge seed (AAS) and garden isolates. The Pseudomonas group, were reported as strict
68
aerobes and are known for the presence of a variety of method is orgnosolv pretreatment, and the liquid hot water
metabolic pathways that are capable of degrading complex pretreatment is slightly better than diluted sulfuric acid
organic compounds including xenobiotics [5]. The genera pretreatment for reducing sugars production. The reducing
of Bacilli were ubiquitous among the mixture of microbial sugars yield produced after optimization is much higher
community. In Bacillus sp., xylanase is induced by xylose, than those of enzymatic hydrolysis in reported in prior
but is repressed in the presence of glucose [6]. Three literature [8] [9].
fungi isolates from the microbial community in this study
are assigned to Pseudallescheria sp. which is known as a
potential animal pathogen and can degrade a wide range
of organic waste [7].
Figure 2. Reducing sugars yield by the best

microbial “cocktail”.
Figure 1. The gram staining pictures of the players

of the best combination.
To enhance reducing sugars yield, seven pure microbial

strains consisting of Microbacterium sp. F28 (M1),
Cellulosimicrobium sp. C10 (M2), Tsukamurella sp. C35
(M3), Pseudallescheria sp. D42 (M4), Bacillus sp. F4 (M5),
Strenotrophomonas sp. B15b (M6) and Pseudomonas sp.E8
(M7) were combined to investigate the best defined microbial
combination. The best defined microbial combination was
Figure 3. Comparison of reducing sugars yield under three
selected, which was the combination of Microbacterium sp. different pretreatment methods at 56.9 ºC, pH5.8, 28.9g initial
F28, Tsukamurella sp. C35, Pseudallescheria sp. D42, and lignocellulosic waste (From left to right). 1. diluted acid
Bacillus sp. F4 (Figure 1). The maximum reducing sugars pretreatment, 2. liquid hot water pretreatment,
yield was 1,653 mg/L (Figure 2), which is ten times more 3. orgnosolv pretreatment.
than the original consortia. Bacillus sp. F4 is a critical
factor. Pseudallescheria sp. D42, which is the only fungus CONCLUSIONS
in the combination, was the second most important factor
to obtaining more reducing sugars.
The results of this study indicated that microbial hydrolysis

can effectively convert lignocellulosic waste to reducing
Optimization experiments were conducted to investigate sugars and it is an alternative hydrolysis method for fuel
the effects of substrate concentration, pH and temperature ethanol production.
under three types of pretreatment methods by the defined
microbial “cocktail” selected in above section. The
maximum reducing sugars production yield of 173 mg
REFERENCES
reducing sugars/g lignocellulose was estimated at the
optimum conditions of temperature 56.9 ºC, pH 5.8, and
[1] Mabee, W.E. and J.N. Saddler, 2010. “Bioethanol from
lignocellulose concentration 28.9 g/L under orgnosolv lignocellulosics: Status and perspectives in Canada”.
pretreatment (Figure 3). The important degree of three Bioresource Technology, 101(13): pp. 4806-4813.
variables on reducing sugars yield is: pH > temperature
[2] Wenzel, M., et al., 2002. “Aerobic and facultatively anaerobic
> lignocellulose concentration. The best pretreatment cellulolytic bacteria from the gut of the termite Zootermopsis
69
angusticollis”. Journal of Applied Microbiology, 92(1): pp. [6] Beg, Q.K., et al., 2001. “Microbial xylanases and their
32-40. industrial applications: a review”. Applied Microbiology and
[3] DuBois, M., et al., 1956. “Colorimetric method for Biotechnology, 56(3): pp. 326-338.
determination of sugars and related substances”. Analytical [7] Anastasi, A., G.C. Varese, and F.M. Valeria, 2005. “Isolation
Chemistry, 28(3): pp. 350-356. and identification of fungal communities in compost and
[4] Miller, G.L., 1959. “Use of dinitrosalicylic acid reagent for vermicompost”. Mycologia, 97(1): pp. 33-44.
determination of reducing sugar”. Analytical Chemistry, 31(3): [8] Saha, B.C. and M.A. Cotta, 2008. “Lime pretreatment,
pp. 426-428. enzymatic saccharification and fermentation of rice hulls to
[5] Landy, E.T., et al., 2008. “Bacterial diversity associated with ethanol”. Biomass and Bioenergy, 32(10): pp. 971-977.
archaeological waterlogged wood: Ribosomal RNA clone [9] Jensen, J.R., et al., 2010. “Effects of dilute acid pretreatment
libraries and denaturing gradient gel electrophoresis (DGGE)”. conditions on enzymatic hydrolysis monomer and oligomer
International Biodeterioration & Biodegradation, 61(1): pp. sugar yields for aspen, balsam, and switchgrass”. Bioresource
106-116. Technology, 101(7): pp. 2317-2325.
70
REMOVAL OF PHARMACEUTICAL
COMPOUNDS IN TROPICAL
CONSTRUCTED WETLANDS
Dongqing Zhang (dqzhang@ntu.edu.sg)
Sara Sadreddini (S.Sadreddini@ntu.edu.sg)
Junfei Zhu (JFZhu@ntu.edu.sg)
Nguyen Anh Tuan (ATNGUYEN@ntu.edu.sg)
Richard. M. Gersberg (rgersbe@mail.sdsu.edu)
Soon Keat Tan (ctansk@ntu.edu.sg)
ABSTRACT: The ability of a tropical horizontal subsurface constructed wetlands (HSSF CWs) planted with Typha Angustifolia to remove
four widely used pharmaceutical compounds (carbamazepine, declofenac, ibuprofen and naproxen) at the relatively short hydraulic residence
time of 2 to 4 days was documented. For both ibuprofen and naproxen, pharmaceutical compounds with low Kow values, the planted
beds showed significant (p<0.05) enhancement of removal efficiencies (80% and 91%, respectively at the 4 day HRT) as compared to
unplanted beds (60% and 52%, respectively). The presence of plants resulted in the removal of these pharmaceutical compounds from
artificial wastewater. Carbamazepine, considered as one of the most recalcitrant pharmaceuticals, and declofenac, showed low removal
efficiencies in our CW, and is attributable to their higher hydrophobicity. The fact that the removal of these compounds could be explained
by the sorption onto the available organic surfaces, explains why there was no significant difference (p>0.05) in their removal efficiencies
between planted as compared to unplanted beds. No statistical significant differences (p>0.05) were observed for the removal efficiencies
of any of the pharmaceuticals tested for the 2-day HRT as compared to that corresponding to 4-day HRT. The rather efficient removal
shown by the tropical wetlands with HRTs of 2 to 4 days indicates the possibility of using such a CW system (with less land/footprint
requirements) for removing certain pharmaceutical compounds from drinking water reservoirs.
INTRODUCTION al., 2009) in engineered low impact systems such as CWs

(Matamoros et al., 2005).
Many pharmaceutical compounds are now considered as
emerging contaminants of environmental concern because of In this paper, we focused on the removal efficiency of four
their widespread use, continuous release, their persistence, pharmaceuticals: carbamazepine, declofenac, ibuprofen and
and increasing evidence of their ecotoxicological (if not naproxen, in a HSSF CWs in a tropical environment. These
human health) effects (Buser et al., 1999). Since some pharmaceuticals were chosen because they were widely
pharmaceutical compounds are not completely removed used and reportedly present in the effluents of wastewater
by conventional wastewater treatment, they are ubiquitous treatment plant (WWTP) effluents. The objectives of
and persistent pollutants in receiving waters worldwide, this study were to (i) compare the removal efficiency of
especially where municipal wastewaters are discharged selected pharmaceuticals in a HSSF CWs planted with
into waterways (Ellis et al., 2006). Typha Angustifolia and in unplanted bed (sand filter); (ii)
compare the removal efficiency of selected pharmaceuticals
Technologies do exist which can lower the level of at relatively low hydraulic residence time (2 and 4 days),
pharmaceuticals discharged into receiving waters (e.g., and (iii) determine quantitatively the role of aquatic plant
ozonation, reverse osmosis, and advanced oxidation plays in removal of the selected pharmaceuticals in a
processes, see Ternes et al., 2002). Such treatment is tropical CW.
extremely expensive (Heberer, 2002). Therefore use of

constructed wetlands (CWs) is growing in popularity as
a low impact and economical alternative for purifying METHODS AND MATERIAL
contaminated waters. Wetlands ecosystems contain a rich
biological diversity and contribute great benefits to society Description of constructed wetlands
by recharging aquifer, retaining sediments and nutrients,
controlling floods and microclimate stabilization (Mitsch The wetland beds were 120 cm long, 60 cm wide and 60
et al., 2008). The complexity of such processes makes it cm deep. Thickness of gravel bed was 0.30m with 4-10 mm
difficult to ascertain the primary removal mechanism for gravel (D60=3.5mm). The porosity was 0.45. Three HSSF
each class of contaminants. While there were extensive CWs were planted with cattail (Typha Angustifolia) at a
research on CWs removal of organic matters (Mitsch et density of 9-10 plant/m2, and 3 beds without plant were
al., 2005), relatively few work has been conducted to used as “unplanted” control beds. All the containers had
evaluate pharmaceutical removal efficiencies (Llorens et a flat bottom and a horizontal drainage pipe 0.4 m long
71
and 50 mm in diameter located at the lower edge of the fibre membrane filter (Millipore, USA) and then acidified
containers. Water depth was maintained at 5 cm below gravel to pH 2 with hydrochloric acid. The SPE cartridges were
surface and the hydraulic loading rate was maintained at conditioned using 5 ml n-hexane, 5ml ethyl acetate, 10 ml
2.96 cm/day. Different from a conventional continuous flow methanol and 10 ml of Milli-Q water (pH=2) at a flow-
wetland, our design was a batch-flow constructed wetland. rate of approximate 3 ml/min. Samples were percolated to
Artificial wastewater enriched with selected pharmaceutical the SPE cartridges through a Teflon tube at a flow-rate of
compounds was rapidly filled in each bed and then drained approximate 10 ml/min. The cartridge filter was then rinsed
completely every 2 or 4 days. with methanol: milli-Q water (pH=2) = 10:90, followed by
20 ml of Milli-Q water (pH=2). Thereafter, the cartridges
were allowed to dry for 30 min and then eluted with 5
General parameter detection ml ethyl acetate with elution solutions collected in 15 ml
calibrated centrifuge tubes. The extracted solution was then
This experiment was carried out in April 2010. During concentrated to ca. 400 μl under a gentle nitrogen stream
the experiments, all the beds were fed with synthetic and was then reconstituted to 500 μl with methanol.
wastewater with the same organic load. Effluent samples
were collected from each bed every two and four days in
a 1-L amber glass bottles. The samples were immediately HPLC Liquid chromatography analysis for
analyzed to determine the reduction in concentration of pharmaceutical
the general parameters, i.e., COD, ammonia-N (NH4-
N), nitrate (NO3) and total phosphorus (TP). General Chromatographic analysis was performed on a Shimadzu
parameters were analyzed by using spectrophotometer Ultra Fast Liquid Chromatograph (UFLC) (Shimadzu,
(HACH-DR3800, USA) in accordance with the conventional Japan) equipped with a quaternary LC-20AD pump, a
methods (Standard Methods for Examination of Water and CTO-20A oven, and a SPD-M20A Diode Array Detector
Wastewater - APHA, 1989). Total organic carbon (TOC) was (DAD). The injector was SIL-20A HT fitted to a Shimadzu
conducted by using Total Organic Carbon Analyzer (TOC- autosampler with a 20 μl sample loop. Chromatographic
Vcsh, Shimadzu, Japan). Dissolved oxygen (DO), pH value separation were carried out using a GracePureTM SPE C18-
and conductivity were measured by using Multi-Parameter Max (4.6*150 mm, 5 μm) cartridge column protected by
Digital Meter (HACH – HQ40d, USA) directly. a ODS-3 (C18) (4.6*50 mm., 5 μm) guard column (Alpha
Analytical). The system was controlled using an interface
module and a personal computer. Chromatograms were
Injection of pharmaceutical products and chemical for processed using a “Shimadzu LCSolution program”.
pharmaceutical dectection
HPLC-grade acetonitrile, methanol, Hexane and ethyl RESULTS AND DISSCUSSION

acetate (both HPLC grade) were obtained from Fisher
(USA). Ultrapure water was obtained from Milli-Q Results
water purification system (Millipore, Bedford, USA).
Carbamazepine, declofenac, ibuprofen, and naproxen (97- Table 1 presents the mean levels of pharmaceutical
100% purity) were obtained from Sigma-Aldrich. The compounds, and their removal efficiencies in the CWs.
microfiters of solid phase extraction (SPE) cartridges,
packed with GracePureTM C18 – Max SPE column 500 Table 1. Mean levels of pharmaceutical compounds, and
mg, 6 mg, were purchased from Belgium. The 0.45 um their removal efficiencies in the CWs
glass fiber filters of 47 mm (Whatman) were purchased Cabamaze
from Schleicher & Schuell (Germany). Stock solution of Naproxen Diclofenac Ibuprofen
pine
(μg1–1) (μg1–1) (μg1–1)
500 μg/ml was prepared in methanol and stored at 4 °C. (μg1–1)
Working solutions were prepared by diluting the stock 2-day HRT
standard solution with methanol. HFCW 17.9±2.5 4.3±1.8 12.8±2.8 7.3±3.4
Removal 28.4±10.38 2.8±7.1* 47.5±8.1 71.0±15.5

SF 17.8±2.8 13.1±3.7 14.2±3.8 11.9±6.6
With respect to the pharmaceutical injection, 50 L of Removal 28.8±11.3 49.5±13.0* 46.7±12.3 56.6±24.6
synthetic wastewater were spiked with 1.25 mg of each 4-day HRT
pharmaceutical compound to obtain a final concentration HFCW 17.0±4.7 2.6±2.4 11.3±2.8 4.3±2.1
of 25μg L-1. The injection experiment lasted 60 days. Removal 17.9±2.0 12.0±4.2 14.6±2.8 10.1±5.7
Removal 28.3±8.1 52.0±17.3* 41.1±11.3* 59.8±22.7*
Solid phase extraction (SPE) The removal efficiency of ibuprofen in our experiment
was higher in the planted beds (71.0% for 2-day HRT and
All the effluent samples were kept refrigerated at 4°C and 79.7% for 4-day HRT) than that for the unplanted beds
analyzed within 24 hours. Prior to extraction, 500 ml of (56.7% for 2-day HRT and 60.0% for 4-day HRT), but
effluent wastewater were filtered through a 0.45 μm glass the enhancement by plants was only significant (p<0.05)
at the 4-day HRT.
72
A significant difference (p<0.05) was observed for naproxen plated CWs, which can promote aerobic reactions leading
removal efficiency between the planted HSSF beds (82.8% to higher removal efficiency. It is also known that root
for 2-day HRT and 91.3% for 4-day HRT) and the unplanted exudates released by the plant in the rhizosphere, are
beds (49.5% for 2-day HRT and 51.8% for 4-day HRT) at known to result in intense microbial activity in the vicinity
both HRTs (Table 1). of roots (Brimecombe et al., 2001). The establishment of
large numbers of metabolically active populations of soil
As for carbamazepine, no significant differences in microbes in the rhizosphere is certainly important (Brix,
carbamazepine removal efficiency between planted beds 1997), as the microbial population found in the soil is
(28.4% for 2-day HRT and 26.7% for 4-day HRT) and associated with the plant roots, which can reach up to
unplanted beds (28.8% for 2-day and 28.3% for 4-day 109 to 1012 per gram of soil (Whipps, 1990). In addition,
HRT) were observed in our study. Another pharmaceutical the possibility that root exudates also may play a role in
compound, diclofenac has also been reported as a recalcitrant induction of specific metabolic activities conferring the
compound in microcosm experiments, membrane bioreactor ability to degrade certain pharmaceuticals, or increase
systems, and activated sludge STP (Hebere, 2002), and bioavailability of pharmaceuticals by acting as surfactants
in our study the removal efficiency of diclofenac in the or transporters, should not be overlooked.
planted beds ranged from 47.5 to 55.4%, compared to that
of unplanted bed of 41.1% to 46.7% (Table 1). Carbamazepine is considered to be one of the most
recalcitrant pharmaceuticals and the removal behavior of
such compounds is completely different from the others
3.2 Discussion above. The recalcitrant nature of this substance has also
been previously reported at other WWTPS. Its low removal
In our present study, ibuprofen and naproxen, both efficiency can be attributed to its higher hydrophobicity,
pharmaceutical compounds with low Kow values, showed and the major fraction of removal of this compound could
significantly (p<0.05) higher removal efficiencies in the be explained by the sorption onto the available organic
planted beds as compared to that in the unplanted beds surfaces (Matamoros et al., 2005). Surprisingly, comparing
(Figure 2). This finding is also consistent with those with other unplanted beds, HSSF CWs, VSSF CWs, or
reported by other researchers, which indicated the removal even WWTPs, the removal efficiencies of cabamazepine in
efficiency for ibuprofen in planted HSSF (with sand depth our study shows much better outcome. However, Hijosa-
of 0.27 m) ranged from 71-80% and 80-90% for naproxen Valsero et al (2010) reported that carbamazepine removal
(Matamoros and Bayona, 2006), but only 49-90% for was favored by plant presence, which is not consistent
ibuprofen and 66-80% for naproxen in unplanted beds with our results.
(Matamoros et al, 2007). This may be well attributed to
the rhizosphere effect, since it has been extensively shown As for diclofenac removal, the sorption of these compounds
that rhizosphere aeration plays an important role part in onto organic matter retained in the gravel bed is an
the establishment of an oxidizing environment to support important removal mechanism due to their hydrophobic
high microbial activity (Reddy et al., 1989). structure, which could be ascribed to specific structural
characteristics. Surprisingly, a significant difference between
Surprisingly in our CW system, levels of DO, COD and planted and unplanted bed was observed for planted beds as
TOC in the planted beds as compared to those in the compared to the unplanted beds but only at a 4-day HRT
unplanted beds were found to be not statistically different (Table 1). Further investigation of its removal pathway
(p>0.05). Since the levels of the pharmaceutical compounds and mechanisms is needed to be carried out in future
enriched into our wetlands were in the μg/L range (as experiments.
opposed to COD levels in the tens to hundreds of mg/L
range), we would have expected that the oxygen demand
to satisfy the requirements for pharmaceutical removal, CONCLUSIONS
would be relatively small. Therefore, our results indicate
the possibility of some rhizosphere effect, aside from This study demonstrated that CWs can be cost-effective
rhizosphere aeration alone, as playing a significant role and sustainable alternative for removing selected emerging
in the efficient pharmaceutical removal in tropical CWs contaminants. The key results can be drawn as follows:
we observed.
CWs can offer comparable or even better pharmaceutical
The removal of ibuprofen and naproxen that were observed removal efficiencies as compared to conventional WWTPs.
in planted beds was also reported by Hijosa-Valsero et al Both ibuprofen and naproxen, pharmaceutical compounds
(2010). The author also indicates that ibuprofen does not to both with low Kow values showed significant (p<0.05)
bind significantly to organic matter retained in the gravel enhancement of removal in planted beds as compared to
beds or pond sediment, and an HFCW planted with T. unplanted ones. The presence of plants seems to favor
latifolia played a significant role in the removal of ibuprofen. the removal of certain pharmaceuticals from wastewater.
This is attributed to the effect of rhizosphere aeration and Rhizosphere aeration plays an important role part in the
more oxidized conditions in these shallow subsurface of establishment of an oxidizing environment, and the more
73
oxidized condition in the planted beds can promote aerobic [5] Heberer, T., 2002. “Occurrence, fate and removal of
reactions to support high microbial activity, leading to higher pharmaceutical residues in the aquatic environment: a review
degree of biodegradation and removal efficiency. of recent research data”. Toxico. Lett. 2002, 31, 5-17.
[6] Hijosa-Valsero, M., Matamoros, V., Sidrach-Cardona, R.,
Carbamazepine, considered as the most recalcitrant Martín-Villacorta, J., Bécares, E. and Bayona, J.M., 2010.
pharmaceuticals, and declofenac showed low removal “Comprehensive assessment of the design configuration of
constructed wetlands for the removal of pharmaceuticals and
efficiencies in our CW, attributed to their higher
personal care products from urban wastewaters”. Water Res.
hydrophobicity. The fact that the removal of these 44 (2010) 3669-3678.
compounds could be explained by the sorption onto the
[7] Llorens, E., Matamoros, V., Domingo, V., Bayona, J.M.
available organic surfaces, explains why there was no and García, J., 2009. “Water quality improvement in a full-
significant difference (p>0.05) in their removal efficiencies scale tertiary constructed wetland: Effects on conventional
between planted and unplanted beds. and specific organic contaminants”. Science of the Total
Environment, 407 (2009) 2517-2524.
No statistical significant differences (p>0.05) were observed [8] Matamoros, V., García, J. and Bayona, J.M., 2005. “Behavior
for the removal efficiencies of any of the pharmaceuticals of selected pharmaceuticals in subsurface flow constructed
tested for a 2-day HRT as compared to a 4-day HRT. wetlands: a pilot-scale study”. Environ. Sci. Technol. 2005,
However, the rather efficient removal shown by the tropical 39, 5449-5454.
wetlands with HRTs of 2 to 4 days indicates the possibility [9] Matamoros, V. and Bayona, J.M., 2006. “Elimination of
of using such a CW system (with less land requirements) pharmaceuticals and personal care products in subsurface
for removing certain pharmaceuticals from drinking water flow constructed wetlands”. Environ. Sci. Technol. 2006, 40,
reservoirs in Singapore or other tropical regions. 5811-5816.
[10] Matamoros, V., Arias, C., Brix, H. and Bayona, J. M., 2007.
“Removal of pharmaceuticals and personal care products
(PPCPs) from urban wastewater in a pilot vertical flow
REFERENCES
constructed wetland and a sand filter”. Environ. Sci. Technol.
2007, 41 8171-8177.
[1] Brimecombe, M.J., Leij, F. A.A.M. and Lynch, J.M., 2001.
“Rhizodeposition and microbial populations”. In: The [11] Mitsch, W. J., Day, J. W., Zhang, J. L. and Lane, R., 2005.
rhizosphere: biochemistry and organic substances at the “Nitrate-nitrogen retention by wetlands in the Mississippi
soil-plant interface, Pinton, R.,Varanini, Z., Nannipieri, P., River Basin”. Ecological Engineering, 24: 267-278.
Eds., Marcel Dekker, New York, 2001, pp. 74-98. [12] Mitsch, W.J., Tejada, J., Nahlik, A., Kohlmann, B., Bernal,
[2] Brix, H., 1997. “Do macrophytes play a role in constructed B. and Hernández, C.E., 2008. “Tropical wetlands for climate
treatment wetlands?” Wat. Sci. Tech., Vol. 35, No. 5, pp 11- change research, water quality management and conservation
17. education on a university campus in Gosta Rita”. Ecological
Engineering, 34 (2008) 276-288
[3] Buser, H., Poiger, T. and Mueller, M. D., 1999. “Occurrence
and environmental behaviour of the chiral pharmaceutical drug [13] Ternes, T.A., Stüber, J., Herrmann, N., McDowell, D., Ried,
ibuprofen in surface waters and in wastewater”. Environ. Sci. A. and Kampmann, M., 2003. “Ozonation: a tool for removal
Technol. 1999, 33, 2529-2535. of pharmaceuticals, contrast media and musk fragrances from
wastewater?” Water Res 2003, 37(8): 1976-82.
[4] Ellis, J. B., 2006. “Pharmaceutical and personal care products
(PPCPs) in urban receiving waters”. Environ. Pollut. 144, [14] Whipps, J.M., 1990. “Carbon economy”. In J.M., Lynch, ed,
184-189. The Rhizosphere. Wiley, New York, pp 59-97.
74
RESPONSES OF FLOATING BREAKWATER

TO REGULAR WAVES
Wenbin Zhang (wbzhang@ntu.edu.sg)
INTRODUCTION
With the development of a large number of small marinas

and recreational harbors, many new types of breakwaters
have been proposed to reduce the energy transmitted into the
harbors. However, in deep waters, traditional bottom-fixed
breakwaters may lead to more expensive construction costs.
Recently, floating breakwaters are considered as an efficient
alternative approach to protecting marinas and harbors
from coastal waves. In this report, an experimental study
on motion responses and wave scattering of pontoon-type
floating breakwater is reported, aiming at collecting a set Figure 1. A view of the breakwater model in the wave flume.
of reliable experimental data under various wave conditions
for future theoretical and numerical studies. A mooring cable with the anchor is shown in the left panel
of Figure 2.
EXPERIMENTAL SETUP
The experiments were conducted in a water flume located

in the Hydraulics Laboratory at Nanyang Technological
University (NTU), Singapore. The concrete-walled wave
flume used in the experiments is about 45 meters in total
length, 1.55m in width and 1.55m in depth. At one end of Figure 2. Anchor with mooring cable, ball-bearing
the flume, there is an adjustable slope beach covered with structure, and mounted IMU.
several porous pads to minimize wave reflections.
In order to reduce the potential risk of damaging the model
A piston-type wave generator is equipped with a system
and two walls of wave flume and maintain a two dimensional
so-called Active Wave Absorption Control System
motion, ball bearing structures were designed and fixed at
(AWACS), which was developed by Danish Hydraulic
the two sides of the floating breakwater. The ball bearings
Institute (DHI). The principle used in the AWACS is to
can prevent the floating structure from colliding with the
measure the (reflected) waves on the wave paddle, and
walls of the wave flume when the floating breakwater
by means of digital filters and the servo system, to isolate
responses to the regular waves. The ball bearing structure
the reflected part and then impose an opposite movement
is shown in the middle panel of Figure 2.
of the wave paddle, which results in the desired incident
wave conditions.
The responses of the floating breakwater model to waves
were recorded by an Inertia Measurement Unit (IMU), which
The breakwater model used in this study is 1.420m long,
can measure the accelerations of three translational motions
0.75m wide and 0.410m high. The total weight of the

and angular velocities of the three rotational motions. The
breakwater reaches 230kg. The draft was set at 0.217m in
IMU was mounted at one corner on the top plate of the
our experiments. The normalized natural frequencies of three
model, referring to the right panel of Figure 2. The IMU
modes were found to be 0.89. A view of the breakwater
sampling frequency used in this study was 200Hz.
model installed in the wave flume is shown in Figure 1.
The experimental configuration of floating breakwater is
The mooring system consists of 2 components: mooring
illustrated in Figure 3. A total of 5 wave gauges were
cables and anchors. It should be noted that the anchors
used for recording wave elevations. There were three wave
need proper weights and volumes so that they can provide
probes placed in front of the model to separate the incident
sufficient tensional forces for the mooring cables, but do
waves from reflected waves. The other two probes were
not disturb the water waves significantly. Stainless steel
placed after the model to record the transmitted waves.
chains were used as they are strong enough and do not
The distances between the three wave probes in front of
rust when being placed in the wave tank for a long time.
75
the model were chosen based on the requirement of the

two-point wave separation method (C.R. Liu, et al. 2009;
Goda, Y. 2000). On each testing day, pre-calibration and
post-calibration were performed to ensure the accuracy of
the experimental results.
Figure 3. Experimental configuration of breakwater.

Figure 5. Measured RAOs for pitch, surge &
heave of the floating breakwater.
CONCLUSIONS
Figure 4 shows the variation of measured reflection
coefficients and transmission coefficients with normalized
A series of experiments have been conducted to investigate
wave frequency parameter, i.e. ω2B/2g, where ω is circular
the wave scattering and motion responses of a 2-dimensional
frequency of the incident waves, and B is the breadth of
floating breakwater under regular waves. The influences of
floating breakwater.
wave period (ranging from 0.8s to1.6s) on the reflection
coefficient and transmission coefficient were studied. The
RAOs were thoroughly analyzed. Our experimental results
agree with the existing results found in the literature
(Sannasiraj, S.A et al. 1998; Yamamoto, T. 1981), indicating
that our data are reliable. These data obtained are intended
for verifying theoretical and numerical models in future
studies. It is also our motivation that the results presented
here will be a foundation for our further study on the effects
of different damping devices on the responses of floating
breakwaters to regular and irregular waves.
REFERENCES
Figure 4. Measured reflection coefficient (Cr) & transmission [1] C.R. Liu, Z.H. Huang, and S.K. Tan, 2009. “Nonlinear
coefficient (Ct) of the floating breakwater. scattering of non-breaking waves by a submerged horizontal
plate Experiments and simulations”. Ocean Engineering, 36
(2009) 1332-1345.
The measured Response Amplitude Operators (RAOs) for [2] Goda, Y., 2000. “Random Seas and Design of Maritime
pitch, surge and heave responses are shown in Figure 5 as Structures”. 2nd Edition, World Scientific, Singapore.
functions of the normalized wave frequency parameter. It [3] Sannasiraj, S.A., Sundar, V. and Sundaravadivelu, R., 1998.
can be found that the maximum values of RAOs for three “Mooring forces and motion response of pontoon-type floating
modes of motion occur when the incident wave frequency breakwaters”. Ocean Engineering, 25(1): 27-48.
equals to the natural frequencies of the breakwater model.
[4] Yamamoto, T., 1981. “Moored floating breakwater response

This can be attributed to the resonance phenomena of the to regular and irregular waves”. Applied Ocean Research, 3,
floating breakwater. 114-123.
76
SOLITARY WAVE INTERACTION WITH A

SLOTTED BARRIER: WAVE SCATTERING
AND HYDRODYNAMIC FORCES
Zhida Yuan (yuan0025@ntu.edu.sg)
ABSTRACT: Slotted barrier are low cost structures that can be very effective in reducing the transmitted energy of long waves. In this
study, the transmission and reflection of tsunami waves, with the leading wave being modelled by a solitary wave, through slotted barriers
in the form of a row of circular cylinders were studied experimentally. The results were also analyzed by a method based on long wave
approximations. It is found that the spacing between two adjacent cylinders is one of the main factors that control the transmission of
solitary waves through slotted barriers. Hydrodynamic forces induced by solitary wave were measured and drag force coefficients of
the slotted barrier are discussed in this paper.
INTRODUCTION EXPERIMENTAL SET-UP
Tsunami waves generated by mighty underwater earthquakes/ In laboratory simulation, the first peak of tsunami waves
landslides, which can occur at any time, can strike in is normally modeled by a solitary wave in view of
minutes, and cause damages to coastal areas. Active the extremely long length of such waves. A series of
protective measures such as breakwaters are also necessary experiments were conducted in a wave flume located at
to prevent ships from breaking mooring lines and hitting the Hydraulics Laboratory, NTU, Singapore, to study the
the port facilities because of the tsunami-induced current. transmission/reflection and wave forces of solitary waves
through pile/slotted breakwaters consisting of an array of
Pile or slotted breakwaters are low cost breakwaters (see
circular cylinders of diameter D = 3 cm. Two wave probes
Mani & Jayakumar (1995) for cost estimation for pipe/ were used to measure the surface elevations at location G1
slotted breakwaters) that can be very effective in reducing and G2, and the middle cylinder was instrumented with
the transmitted energy of long waves (see Mei et al. a 3-D force transducer at its top end to measure the total
(1974)). Figure 1 shows a section of pile breakwater along wave forces acting on it. Figure 2 shows a view of the
Singapore coast. A lot of research has been carried out on pile breakwater used in our experiments. The wave flume
the interactions of regular waves with slotted barriers in the was 32 m long and 54 cm wide. Installed at one end of
absence of currents (e.g. Kakuno & Liu (1993); Isaacson et the wave flume was a piston- type wave generator (HR
al. (1998); Huang (2007)). Recently, the effects of currents Wallingford), which was used to generate the solitary waves.
on the scattering of regular waves by slotted barriers were Figure 3 shows the experimental setup, where the wave
examined by Huang (2006) and Huang & Ghidaoui (2007). probe G1 was used to measure the incident and reflected
However, few studies about tsunami wave interaction with waves while the wave probe G2 was used to measure the
slotted barriers are reported. transmitted wave. Civil Engineering Research • January 2011
Figure 2. A view of the slotted barrier installed

in the wave flume.
Figure 1. A segment of pile breakwater found in a section of
the Singapore coast.
The main objective of this study is to examine experimentally

the transmission and reflection of tsunami waves (solitary
waves) through pile breakwaters; moreover, wave forces
were measured and hydrodynamic coefficients were
investigated and compared. Figure 3. Sketch of experimental setup showing the relative
locations of the slotted barrier and the two wave probes.
77
To study the effects of water depth and incident wave For a given incident wave height HI/h, the measured
height, five wave heights varying from 4 cm to 8 cm were transmission coefficient increases with increasing spacing-
examined for h = 15 cm and six wave heights varying to diameter ratio S/D, while the reflection coefficient
from 5 cm to 10 cm were examined for h = 20cm. Three decreases with increasing S/D. These trends agree with
different slotted barriers were used in the experiments, the following theoretical observations: in the limit of
with the center to-center distance between two adjacent S / D → ∞, theoretically CT → 1 and CR → 0; in the
limit of S / D → 0, theoretically CT → 0 and CR → 1.
cylinders (spacing) being S = 4.5 cm, S = 4.2 cm, S =
The transmission coefficients are calculated by long wave
3.64 cm, respectively. As the diameter of the cylinder was theory by Huang & Ghidaoui (2007) for a given barrier
3 cm, the spacing-to-diameter ratio S/D ranged from 1.50 and incident wave conditions. Numerical experiments
to 1.21 in this study. show that the following values of f can produce the best
fits to the measured transmission coefficients for the three
barriers: f =3 for S/D =1.5, f =6 for S/D =1.4 and f = 18
RESULTS AND DISCUSSIONS for S/D = 1.21. Figures 4~6 show the comparison between
the predicted and measured transmission coefficients for all
We can define the reflection (CR) and transmission (CT) three barriers studied in the experiments. It can be seen that
the long wave approximation can predict the transmission
coefficients by
coefficient satisfactorily.
HR H
CR = , CT = T …(1)
HI HI
where HR, HT and HI are the heights of reflected, transmitted
and incident solitary waves, respectively. These coefficients
are determined both experimentally and by using the
theory developed for long waves interacting with a slotted
barrier in the presence of a uniform current (Huang &
Ghidaoui (2007)).In our experiments, two water depths
were examined with D/h = 1/5 and D/h = 3/20, respectively.
After analyzing our experimental data, it was found that
the measured reflection and transmission coefficients are
nearly independent of D/h. Figures 4~6 show the measured
reflection and transmission coefficients, together with those
Figure 5. Comparison between the measured and predicted
predicted by long wave theory of Huang & Ghidaoui hydrodynamic coefficients for S/D = 1.40. Solid lines (f = 6),
(2007), for spacing-to-diameter ratio S/D = 1.21, 1.40, chains (f =4.8), and dashed lines (f =7.2).
and 1.50, respectively. Here, f is an equivalent quadratic (See Huang & Yuan 2010).
loss coefficient to be determined from the experiments.
For a given S/D, the measured transmission coefficient
decreases with increasing HI/h, while the measured reflection The predicted reflection coefficients are also shown in
Figures 4~6. The long wave approximation can still provide
coefficient is insensitive to the change in HI/h. These trends
reasonable prediction of the reflection coefficients for
are similar to those found in long waves scattered by a relatively small HI/h, i.e., the nonlinearity is weak. For large
slotted barrier in the presence of a uniform current (Huang HI/h, long wave approximation over-predicts the reflection
& Ghidaoui (2007)). coefficients for all three barriers. It is expected that the
nonlinear interaction between the incident and reflected
solitary waves cannot be handled by long wave theory of
Huang & Ghidaoui (2007).

hydrodynamic coefficients for S/D=1.21. Solid lines (f =18),
chains (f=14.4), and dashed lines (f =21.6).
(See Huang & Yuan 2010).
hydrodynamic coefficients for S/D = 1.50. Solid lines (f = 3),
chains (f = 2.4), and dashed lines (f = 3.6).
(See Huang & Yuan 2010).
78
Figure 7 shows the experimentally obtained drag force REFERENCES

coefficients It can be seen that the drag coefficients are
remarkably dependent on cylinder spacing. The mean [1] Huang, Z. and Ghidaoui, M.S., 2007. “A model for the
coefficients curves show that coefficient values for S / D scattering of long waves by slotted breakwaters in the presence
= 1.21 are very close to those for S / D = 1.14 within the of currents”. Acta Mechanica Sinica, 23:1-9.
present test range. The mean curve for S / D = 1.5 clearly [2] Huang, Z. and Yuan, Z., 2010. “Transmission of solitary waves
lies below those of S / D = 1.21 and S / D = 1.14. through slotted barriers: A laboratory study with analysis by
a long wave approximation”. Journal of Hydro-environment
Research, 3: 179-185.
[3] Huang, Z., 2006. “An experimental study of wave scattering
by a vertical slotted barrier in the presence of a current”.
Ocean Engineering, 24:717-723.
[4] Huang, Z., 2007. “Wave interaction with one or two rows of
closely-spaced rectangular cylinders”. Ocean Engineering, 34:
1584-1591.
[5] Isaacson, M., Premasirl, S. and Yang, G., 1998. “Wave
interaction with vertical slotted barrier”. Journal of Waterway,
Port, Coastal and Ocean Engineering, 124 (3): 118-126.
[6] Kakuno, S. and Liu, P.L.F., 1993. “Scattering of water waves
by vertical cylinders”. Journal of Waterway, Port, Coastal
and Ocean Engineering, 119 (3):302-322.
Figure 7. General tendency of drag coefficients versus [7] Mani, J.S. and Jayakumar, S., 1995. “Wave transmission by
dimensionless incident wave heights. suspended pipe breakwater”. Journal of Waterway, Harbor,
Coastal and Ocean Engineering, 121 (6): 335-338.
[8] Mei, C.C., Liu, P.L.-F. and Ippen, A.T., 1974. “Quadratic
CONCLUDING REMARKS head loss and scattering of long waves”. Journal of Waterway,
Harbor and Coastal Engineering Division, 99: 209-229.
In this study, the transmission, reflection and drag force [9] Titov, V.V., Rabinovich, A.B., Mofjeld, H.O., Thomson, R.E.
of solitary wave through pile breakwaters were studied and Gonzalez, F.I., 2005. “The global reach of the 26 December
experimentally, and the measured transmission coefficients 2004 Sumatra tsunami”. Science, 309: 2045-2048.
agree well with those calculated by using long wave
approximation. The transmission coefficient decreases
slightly with increasing HI/h, while the reflection coefficient
is not sensitive to the change of HI/h; the drag force
coefficient decreases slightly with increasing HI/h. The
spacing between the adjacent cylinders has significant
influence on the transmission, reflection and drag force
coefficient; reducing the barrier spacing would remarkably
reduce/increase the transmission/reflection and drag force
coefficient.
79
TECHNOLOGIES FOR WATER

SOFTENING: A REVIEW
Fang Wangxi (WXFang@ntu.edu.sg)
Wang Rong (RWang@ntu.edu.sg)
INTRODUCTION exchange easily with divalent calcium and magnesium ions

in the water. As the water passes through the column, the
The term, “hard water”, refers to water with a high mineral hardness ions replace the hydrogen, sodium or potassium
content that primarily consists of calcium and magnesium ions which are released into the water, thus the softer
cations and other divalent or multivalent metal ions. water is achieved. When the resins become saturated with
Hard water is undesirable as the minerals can produce an hardness cations, they gradually lose their effectiveness
unpleasant taste, react with soap anions to decrease the and must be regenerated. The regeneration is normally
cleaning efficiency, induce scaling and corrosion problems implemented by passing a concentrated brine solution such
and cause serious failures in pipelines of boilers, heat as sodium chloride or potassium chloride through them [4].
exchangers and electrical appliances, etc. [1]. Figure 1 Figure 2 illustrates the reaction occurred during a typical
shows the examples of scaling in a faucet and a pipe. ion exchange resin treatment and regeneration process.
Due to greater affinity of the resin for larger multivalent
cations, the reaction shifts from left to right when the hard
water passes though the resins, while in the regeneration
process, the sodium concentration in the brine solution is
so high that the equilibrium of the reaction shifts toward
left, leading to the replacement of the calcium ions on the
resin with sodium ions.
(a) (b)
Figure 1. (a) Drop coming out of a faucet coated with calcium
from the hard water; (b) Scale reducing the size
of pipework [2].
Water softening is a water treatment process that serves

the removal of dissolved minerals in hard water. Various Figure 2. Schematic of typical cation exchange reactions.
methods have been applied to remove hard water minerals
from natural, industrial or domestic water sources, which
An advantage of the cation exchange treatment process is
include the applications of ion exchange, demineralization,
that it does not change the pH or alkalinity of the water.
chemical precipitation, distillation and membrane separation
Other advantages include excellent process reliability,
[3]. Currently, with the financial support from EWI,
chemical safety and process stability for long-time
Singapore Membrane Technology Center (SMTC) is
performance. That is why water softening by ion exchange
collaborating with Siemens Water Technologies to develop
has been used worldwide, especially in household and
low-pressure hollow fiber membranes for water softening.
industrial water treatments. However, this process has
This study reviews three major treatment technologies
several major drawbacks. Firstly, total dissolved solids
that are currently used in domestic and industrial water

(TDS) in the treated water do not decrease significantly due
softening applications.
to the release of sodium or potassium. Secondly, the excess
brine solution produced during the regeneration process
may be difficult to treat with. In addition, the resins may
ION EXCHANGE be fouled by some metal cations like oxidation by ferric
ions and thus lose the treatment efficiency.
Water softening by ion exchange resins
Alternatively, a class of minerals called zeolites also
In an ion exchange process, hard water passes through exhibits good ion exchange properties. In fact, zeolites
a column of polymer resins. The polymer resins possess were widely used in earlier water softening [5], and later
negatively charged functional groups, which can adsorb replaced by synthetic ion exchange resins. Compared
and bind metal cations. The univalent hydrogen, sodium or with natural zeolites, ion exchange resins are artificially
potassium ions initially contained in the resins are able to
80
synthesized to achieve better treatment efficiency and to 10.6 to achieve efficient precipitation [7]. If the raw
regeneration ability. water contains a high amount of magnesium, excess lime
must be introduced to raise pH above 11 in order to help
precipitating magnesium hydroxide.
Demineralization
Table 1. Lime softening reactions
If the water needs to have the mineral content entirely Hardness Lime Precipitate
removed, it is passed through a cation exchange column in CO2’ + Ca(OH)2 → CaCO3 + H2O
the hydrogen ion form (H+), followed by an anion exchange
Ca(HCO3)2 + Ca(OH)2 → 2CaCO3 + 2H2O
column in the hydroxide form (OH-) to replace all the
cations and anions, respectively. The two-step ion exchange Mg(HCO3)2 + Ca(OH)2 → CaCO3 + MgCO3 + 2H2O
process is called demineralization [6]. The regeneration MgCO3 + Ca(OH)2 → Mg(OH)2 + CaCO3
process involves the recharging of both hydrogen with a
strong acid (usually HCl) and hydroxide ions with a strong If non-carbonate hardness is involved, soda ash addition
base (usually NaOH) for the two columns of resins. is needed. Table 2 shows the additional reactions involved
for lime-soda ash softening. The raw water is assumed to
Water demineralization is also widely used for the contain SO42- in addition to bicarbonate ions.
production of high purity water. However, compared with
the ordinary ion exchange resin process, additional hazards Table 2. Additional reactions for lime-soda softening.
are introduced due to the involvement of strong acid and Hardness Lime/soda ash Precipitate
base. That is the main reason why this technology is less MgSO4 + Ca(OH)2 → Mg(OH)2 + CaSO4
commonly applied in household use.
CaSO4 + Na2CO3 → CaCO3 + Na2SO4
CHEMICAL PRECIPITATION Evaluation of lime softening technology
Water softening by chemical precipitation A key feature of lime softening is that both the calcium in
the raw water as well as the calcium added with the lime
Water softening by chemical precipitation has been applied are precipitated. As a result, the total dissolved solids (TDS)
commonly for large-scale industrial and potable water decrease after the treatment. This is in contrast to the ion
production. There are two major types of precipitation exchange softening where sodium is exchanged for calcium
reactions involved in water softening, which include and magnesium ions, and no significant change occurs in
precipitation by the addition of lime (calcium hydroxide, the level of TDS. The cost-effectiveness in treating large
Ca(OH)2) or soda ash (sodium carbonate, Na2CO3). Lime quantities of surface water also makes lime softening be
is used to remove chemicals that cause carbonate hardness, a major water softening technology in water treatment
while soda ash is introduced to remove chemicals that plant-scale applications [8].
cause non-carbonate hardness. Other chemicals applied
for precipitation treatment include quicklime (CaO) and However, several problems are encountered during the
caustic soda (NaOH). application of this technology. Water after lime treatment
has a high pH, which needs an additional stage of pH
neutralization process. The treatment and disposal of large
Working principles of lime and lime-soda ash amount of high pH sludge is also a problem. Moreover,
treatments continuous removal of the calcium carbonate scale on rapid
mixers and flocculation basin equipment brings additional
During lime softening process, the addition of lime leads operating and maintenance costs.
to the increase of raw water pH and the shift of the
equilibrium of the carbonate species in the water. When pH
gets above 9.5, most of the dissolved carbon dioxide and MEMBRANE FILTRAION
bicarbonate convert into carbonate so that it becomes the
dominant species in the carbonate system. Then calcium Water softening by pressure-driven membrane
carbonate begins to precipitate because the concentrations processes
of the two ions exceed solubility limit of calcium
carbonate. Additionally, magnesium can be precipitated With rapid development of membrane technology in recent
as magnesium hydroxide if excess lime is added into the years, conventional water softening methods involving
system and pH exceeds 11. Table 1 shows the chemical ion exchange resin, zeolites, and lime or lime-soda
reactions occurred during lime softening process. For raw ash treatments are being replaced by membrane-based
water containing minimal magnesium ions, only calcium approaches. Compared with conventional methods, water
needs to be removed. No excess lime needs to be added, softening through membrane filtration does not involve large
and system pH can be maintained in a range from 10.3 quantities of chemicals such as lime and sodium chloride
81
as well as other potentially hazardous chemicals, which scaling and fouling problems and loses performance during
can reduce significantly the TDS of raw water. Therefore, the treatment process [10]. Anti-fouling or pretreatment of
hardness removal through membrane processes has the raw water is therefore required, which reduces the efficiency
potential to offer lower operating and by-product disposal of the overall treatment processes. In addition, the high
costs, increased operation safety and relatively lower energy operating pressure causes high energy consumption. In fact,
consumption [9]. the operating and maintenance costs are a few of the major
drawbacks of RO process in large-scale applications.
Membrane filtration is a separation process with the input
of energy, which serves as a driving force to separate a
mixture. For the application of water softening, reverse Nanofiltration membranes
osmosis (RO) or nanofiltration (NF) membranes are utilized
to physically remove the hard water minerals from the In order to overcome the problems of high energy
raw water source. The type of membrane determines the consumption and membrane scaling and fouling, NF
degree of treatment. membranes are being employed to replace the RO
membranes for certain applications. Since NF membranes
have a less dense structure than RO membranes, a lower
Reverse osmosis applied pressure is required for NF membranes to achieve a
similar water permeation flux as RO membranes. However,
RO has been widely used from household drinking water NF membranes have a nominal pore size in nanometer
purification and industrial water purification to large-scale scale and thus present poorer rejection for univalent ion
water production, seawater desalination and wastewater species like sodium chloride. Nevertheless, NF membranes
treatment. Forward osmosis is the automatic net flow of still remain high rejection to divalent and multivalent ions
water through a semi-permeable membrane from a dilute such as calcium and magnesium ions, which makes it more
solution to a concentrated solution due to the osmotic suitable for water softening.
pressure difference across the membrane. In contrast, RO
is a pressure-driven membrane process which allows water NF membranes are often charged to enhance removal of
in the concentrated solution to pass through the membrane hard water divalent ions from raw water. The separation
and flow into the dilute solution under an external hydraulic mechanisms in this case involve an electrostatic effect
pressure to overcome the osmotic pressure difference. In this named Donnan exclusion and size exclusion. For example,
case, the concentrated solution gets even more concentrated the surface of the NF membrane can carry positive charges.
while more diluted solution i.e. purified water is produced. The divalent ions like Ca2+ and Mg2+ in the feed water
Figure 3 shows a schematic diagram of reverse osmosis will be electrostatically excluded from getting closed to
process. and passing through the membrane. In order to keep the
neutrality of the feed water, the anions like Cl- and SO42-
have to be retained in the solution- thus only water can
pass through the membrane. The Donnan exclusion is less
efficient for the rejection of univalent cations because the
charge density on the univalent ions is smaller than divalent
and multivalent ions [11].
Although the required trans-membrane pressure for

nanofiltration is lower as compared with RO process, the
operating pressure for conventional NF membranes is still
relatively high in order to achieve productive permeation
flux. For water softening application, it is a challenge to
develop NF membranes that remain capable permeation
flux at a low operating pressure so as to reduce energy
Figure 3. Schematic of reverse osmosis process. consumption and membrane fouling tendency. In addition,
NF membranes are mainly in the configurations of flat
RO membranes are able to reject almost all solutes existing sheet and tubular membranes. It is desirable to utilize
in the water such as bacteria, natural organic matters, hollow fiber membranes for large scale water softening
heavy metal ions and multivalent, divalent and univalent processes, as hollow fiber membranes offer better packing
ions. Thus, when using as a means of water softening, density, higher surface area to volume ratio and self-support
RO process does not only remove hard water minerals, capability. To address these challenges, the Singapore
but also purifies the raw water to a higher degree than the Membrane Technology Centre at NTU is currently
requirement of water softening. This characteristic is not conducting research in collaboration with Siemens Water
always an advantage. A higher degree of water treatment Technologies. The project aims to fabricate novel NF hollow
would produce highly concentrated brine that needs to be fiber membranes with high water permeation flux and high
handled properly. The RO membrane also faces serious rejection of divalent ions at a low operating pressure for
water softening.
82
SUMMARY [2] Hustvedt, Drop coming out of a faucet coated with calcium
from the hard water. Retrieved October 21, 2010, from
Three technologies of ion exchange, chemical precipitation http://upload.wikimedia.org/wikipedia/commons/b/bb/Hard_
water_and_drop.jpg
and membrane filtration, which are widely used for water
softening, are reviewed. The pros and cons of each [3] A.P. Sincero and G.A. Sincero, Physical-chemical treatment of
technology are discussed. Ion exchange resin and zeolites water and waste water. IWA Publishing: CRC Press; 2003.
treatment are mainly applied in household and industrial [4] W. Wist, J.H. Lehr and R. McEachern, Water softening
water softening, but the excess brine solution produced with potassium chloride: process, health, and environmental
benefits. J. Wiley; 2009.
during the regeneration process is difficult to handle if the
technology is applied in a large-scale. Lime softening and [5] G.F. Hodkinson, Zeolite water softener, US Patent 1,763,783,
1930.
lime-soda ash treatments are more suitable for large-scale
municipal or industrial water production, while facing [6] C.E. Harland, Ion exchange: theory and practice, Royal
problems involving post-treatment for pH neutralization of Society of Chemistry; 1994
softened water and disposal of large amount of high pH [7] S. Kawamura, Integrated Design and Operation of Water
sludge. These two conventional water softening methods are Treatment Facilities, 2nd ed., J. Wiley; 2000
being replaced by advanced membrane filtration technology, [8] American Water Works Association, American Society of Civil
due to its potential to offer lower operating and by-product Engineers, Water treatment plant design, 3rd ed., McGraw-
Hill; 1998.
disposal costs, and relatively low energy consumption.
Nanofiltration is found to be the most suitable membrane [9] F.E. Duran and G.W. Dunkelberger, A comparison of
process for water softening, but extensive R&D is needed membrane softening on three South Florida groundwaters,
Desalination (1995) 102, 27-34.
to develop high performance NF hollow fiber membranes
for this application. [10] E.M. Vrijenhoek, S. Hong and M. Elimelech, 2001. “Influence
of membrane surface properties on initial rate of colloidal
fouling of reverse osmosis and nanofiltration membranes”.
Journal of Membrane Science, 188, 115-128.
REFERENCES [11] Yaroshchuk, A.E., 2001. “Non-steric mechanisms of
nanofiltration: superposition of Donnan and dielectric
[1 C. Gabrielli, G. Maurin, H. Francy-Chausson, P. Thery, T.T. M. exclusion”. Separation and Purification Technology, 22-23,
Tran and M. Tlili, Electrochemical water softening: principle 143-158.
and application. Desalination (2006) 201, 150–163.
83
TIME-SEQUENCE ANALYSIS OF
JET-FLIPPING OF LOCALIZED SCOUR
BY 2-D WALL JETS
Lim Siow Yong (csylim@ntu.edu.sg),
Xie Chen (xiec0001@ntu.edu.sg)
INTRODUCTION
Severe localized scour downstream of 2-D wall jets may

cause stability problem to the hydraulic structure—it is one
of the major factors to consider when engineers design the
required erosion protection measures. Researchers working
on this issue focused mainly on finding and predicting the
maximum scour depth and other related scour geometries
at the equilibrium condition (Lim and Yu 2002). However,
under certain flow conditions for 2-D jet scour, there would
be jet-flipping, which is a phenomenon where the jet action
Figure 1 Layout of experimental set-up.
would suddenly flip from the bed to the water surface and
vice versa. It would result in a pseudo equilibrium condition.
Video recording of the scouring process showed a scour Table 1 Summary of experimental data.
hole being developed with time when the jet action was No. L t0 dse-dig dse-fill d
digging as a bed-jet. A deep scour hole was formed with (cm) (min) (cm) (cm) (%)
a distinct downstream ridge. This was the digging phase Run16 0 None 14.7 -- --
by the bed-jet. After the hole attained a certain maximum
Run15 15 937 8.1 5.1 37
depth, the jet would suddenly flip to the surface as a
Run13 21 258 6.5 4.6 29
surface-jet. Sand particles on the ridge were seen rolling
down its slope and filling the hole. This is the filling Run14 30 24 3.9 2.8 28
phase by the surface-jet. The hole became shallower with
time under this filling phase. The digging-filling process is the start of filling phase corresponded to when the water
cyclical. In this study, we will present preliminary results surface suddenly transformed from calm to wavy. When
on the cyclical time-sequence of the jet-flipping due to a the water surface changed from wavy to calm again, this
submerged horizontal jet issuing from a sluice gate. moment was recorded as the end of filling or the start
of digging phase. Table 1 summarizes the test conditions
for the 4 runs, where dse-dig is the maximum depth of
EXPERIMENTAL SETUP AND scour recorded during the digging phase, dse-fill is the
minimum depth of scour recorded during the filling phase,
MEASUREMENTS
and d is the % scour depth difference defined as
(dse-dig - dse-fill)/dse-dig, t0 is the average time to complete one
The experiments were conducted in a 8m long, 0.3m wide,
cycle of digging-filling phase, expressed as follows:
0.6m deep glass-walled flume (Fig. 1) in the Hydraulics
n
Laboratory at NTU. A solid Perspex platform was
t0 = ∑(t fi + tdi)
constructed to simulate a rigid apron over which a vertical i=1
n
sluice gate was fixed. The sluice gate can be installed at
any location on the apron. This arrangement facilitates where n = number of digging-filling cycles, tfi and tdi are
the study of the effect of apron length on the jet-flipping the filling and digging time in the ith cycle, respectively.
phenomenon downstream of the gate. Uniform sand with
median grain diameter (d50) of 0.73 mm was used and
the geometric standard deviation (σg) was 1.12. All the Time sequence analysis
experiments ran for 7 days and were set with the same
tailwater depth (Ht = 12.7cm), sluice gate opening (d0 = Under normal circumstances for a typical jet scour
10mm), discharge (Q = 2.128L/s), velocity (u0 = 0.71m/s), experiment, the scour hole will increase with time, and
Froude number (Fr = 2.267), while only the apron length eventually it will attain an equilibrium state after a long
(L) was changed. We used a video to record the scouring scouring time. For the hydraulic conditions used in the
process. From the recording, the duration of the digging present study, at the initial stage the jet would dive and
and filling phases was noted. The end of digging or start to dig at the sediment bed. Soon scour hole of a
84
Figure 2. (a) Bed profile during digging phase by bed-jet with calm water surface (b) Bed profile during filling phase
with wavy water surface by surface-jet. (Arrow shows direction of jet action).
Figure 3. The time-sequence of each digging and filling phase for different apron lengths.
certain depth with a corresponding ridge was formed. The vigorously as the surface-jet is again transformed into a
hole continued to deepen as the scour progressed with time bed-jet. Hence, the process is cyclical with bed-jet scouring
until a stage was reached whence there was no significant the bed during the digging phase and then transforms to a
change in the scour depth, which looked like the equilibrium surface-jet where the hole is back-filled during the filling
state was reached. This period is called the digging phase phase, and vice versa. For a run time of 7 days, we have
of the scour development, and the jet action was generally extracted from the video recording the number of cycles
along the bed and the water surface was calm (Fig. 2a). and their durations for each digging and filling phases (Fig.
However, a short while later, the sediment was observed to 3). Usually, the scour depth reached a maximum in the first
suddenly roll down from the ridge region and back-filling few digging phases and the maximum depth in each digging
into the scour hole. The water surface also became rough phase was almost the same. And the minimum depth in
and wavy. This was caused by the sudden change of the each filling phase was also almost the same.
jet action from a bed-jet to a surface-jet (Fig. 2b). In the
filling phase, the rate at the initial stage of filling was rapid Figure 3 shows that, for each run, the filling time is much
and the scour hole was almost leveled and the ridge was longer than the digging time. The filling time becomes
flattened in the process. For quite a long period during this increasingly longer, while the digging time becomes shorter
filling phase, it seemed an equilibrium state was reached with small scale oscillation over the recording period. For
as there was no significant depth change in the filled scour comparison, a dimensionless time is used (Fig. 4). The
hole. The filling phase usually lasts very long, at the end lines of digging for L = 21cm and 30cm are quite similar.
of which the jet would suddenly begin to scour the bed However, the lines for L = 15cm is very short for the
85
Figure 4. Dimensionless time analysis for three runs with different apron length.
reason that the average time t0 for a shorter apron is much Fig. 5 shows 3 photos at different stages of the scouring
longer, about 937 mins. Although all the experiments ran process for Run 13. It can be seen that the maximum
for 7 days, it is not long enough for Run 15 but too long scour depth did not change, but the ridge height decreases
for Run 14. But the trend can be found that the lines for over the many flipping cycles between Fig. 5a (24/8/10,
Run 15 are quite similar to Run 13 and Run 14. That is, 05:33) to Fig. 5c (30/8/10 16:25). The ridge profiles in
if Run 15 is allowed to run for more days, its trend seems Figs. 5b and 5c can be seen being flattened compared to
to have a similar pattern as Run 13 and Run 14. In Table Fig. 5a, resulting in a reduced ridge height. Fig. 5c shows
1, Run 16 has no apron (i.e. L = 0) and the jet-flipping that the shaded ridge material has been deposited further
phenomenon did not occur. It seems to indicate that the downstream, giving an elongated and flattened bed profile
apron plays an important role for jet-flipping to occur. The downstream of the crest after about 6.5 days of scouring
results in Table 1 show that the average time t0 increases action.
exponentially as the apron length decreases.
As scouring time increases and as more digging-filling
cycles occurred, more particles ‘escaped’ and there are
relatively less particles in the hole-ridge region compared
to earlier cycles. The net effect is that back-filling of
sediments from the ridge to the scour hole takes a longer
time to fill the hole to a level for the next digging action
to be triggered. The digging time remains more or less the
same, and the jet digs to a depth that is required to trigger
the next filling phase. This explains why the maximum
scour depth remains almost constant but the filling time
is increasingly longer as scour progresses.
CONCLUSIONS
The present study mainly observed the jet-flipping

phenomenon. Some conclusions can be drawn from the
study:
(1) Under certain hydraulic conditions, jet-flipping is
observed in localized scour due to a 2-D wall jet.
(2) There are two phases. A digging phase where the jet
behaves as a bed-jet and a scour hole is formed. The
water surface is calm during this phase. After the
scour depth reaches a certain maximum value, the jet
flips from the bed towards the water surface causing
it to become rough and choppy and the filling phase
commences. During this phase, the sediments on the
ridge roll back to fill up the hole until it reaches a
certain minimum scour depth whence the digging
Figure 5. Transition of the scour profile. phase would resume. The digging-filling processes are
cyclical.
86
(3) It was observed that as scouring progresses, the REFERENCES

duration of filling phase increases, while the digging
time remains relatively unchanged. The duration for the [1] Lim, S.Y. and Yu, G., 2002. “Scouring downstream of Sluice
filling phase is much longer compared to the digging Gate”. Proceedings of First International Conference on Scour
phase. The loss of sediment materials on the ridge of Foundations, Texas, USA, 17-20 November 2002, Vol. 1,
as scouring progresses is the main reason for filling pp. 395-409.
becoming increasing longer.
(4) The apron plays an important role in the process of
jet-flipping. The average time t0 which is the time
from the start of a filling phase to the next start of
a filling phase, increases exponentially as the apron
length decreases.
87
INFRASTRUCTURE SYSTEMS AND MARITIME STUDIES
A DECISION SUPPORT SYSTEM FOR

PORT SELECTION
Jasmine Siu Lee LAM (sllam@ntu.edu.sg)
ABSTRACT: The paper presents a web-based decision support system (DSS) for port selection using analytical hierarchy process
(AHP) methodology. AHP is able to assist managers in obtaining a detailed understanding of the criteria and address the port selection
problem utilising multi-criteria analysis. It shows how technology advancement can bring positive effects on strategic planning of
shipping firms.
INTRODUCTION
Users
In liner shipping, service network planning is an important
activity. Selecting the candidate calling ports is the first ♦
step in service network planning. In practice, the current User Interface (Visual Basic)
mode of planning is still to a large extent manual, where
considerable professional knowledge and experience is the
key driver. This may not be an ideal method when time is
of essence and rapid decision making is required to respond Port Selection Module (AHP)
to a dynamic market. Therefore, a DSS that can carry out
the process more efficiently is of utmost importance.
SYSTEM ARCHITECTURE Optimisation
In this paper, we integrate AHP with a DSS using

optimisation system development. To the best of our
knowledge, our work is the first attempt of such an Database server Databases (Access, Excel)
approach. AHP is a multi-objective, multi-criteria theory of
measurement. Once the method is integrated with a DSS,
it can help players in liner shipping industry to shorten Figure 1. Decision support system architecture.
the time needed for making decisions on port selection in
service planning by a more robust approach.
MODEL AND CASE ILLUSTRATION
Difficulties in obtaining data could be a major reason for
hindering the application of choice theories (AHP) on This section illustrates the port selection procedure inside
port selection. However, the proposed DSS is web-based; the DSS. The step-by-step procedure is as follows:
therefore it can be accessed by more users and data 1. Select the regions (or markets) to be covered in the
collection can be carried out faster. Figure 1 shows the service network.
proposed overall structure. Being a web-based system, the
2. Determine the total number of ports to be visited and
DSS includes three tiers. The first tier is the interface that
the number of ports to be visited in region i (xi).
interacts with the user. The second layer is the mechanism
which handles all the calculations and optimisations. 3. Select the port criteria (j = 1, 2,….N) to be considered
The last tier is the database that stores all the necessary and determine the weight of each criterion (wj); the
information for processing. VB.net is used to develop the sum of all selected criteria weights must be equal to
front end of the system, whereas in the second layer, Visual 1. If the user cannot determine the criteria weights
Basic is employed to develop mechanisms to perform directly, the weights are calculated using pair-wise
simple manipulations of the records and interact with the comparison (from AHP using a nine-scale scoring).
database, as well as with the optimisation and financial In pair-wise comparisons, inconsistency may occur.
analysis modules. The database, on the other hand, is mainly Therefore, the consistency ratio (CR) is calculated and
handled by Microsoft software. We use Access and Excel if CR is greater than 0.1, the pair-wise comparison
to maintain all the necessary data. matrix is revised.
88
4. Retrieve scores to all ports in selected regions and The second interface is the comparison and prioritization
criteria from the ports’ database. The score of port k module. The module allows the decision maker to make
under criterion j is Skj. pair-wise comparisons of the selected criteria, and reports
5. The total score of port k (TSk) is the weighted sum the weight of each criterion and the consistency ratio
of the port’s scores in all criteria. (CR) using AHP. In this case, the decision maker focuses
TSk = ∑wjSkj more on container transhipment and its costs than on port
infrastructure. Based on criteria weight calculation in this
6. For each region choose xi ports with the highest case, the weight for port infrastructure, port charge and
score. container traffic is 0.2, 0.4 and 0.4 respectively. CR is 0.00,
7. Perform sensitivity analysis if the decision maker which is less than 0.1 and thus the pair-wise comparison is
wants to know the effects of changing his preference consistent. Same as criteria selection, different users have
for port criteria, due to, for example, a change in port different preferences, and thus the other decision makers
situation. can change the “Input” data (pair-wise comparisons of
the selected criteria) in the module to get the new criteria
weights automatically.
Case study
In the third interface, port data are retrieved from the
The system has been tested and validated with an anonymous database. Thereafter, the ports with the highest overall
liner shipping company in real operation mode to ensure score in each region are selected. In this case, there are
that it is suitable for practical operations. The company is five ports in each region, South Asia and Middle East,
an international liner operator with a wide liner service respectively. According to the requirement for the number
network. In our DSS, there are ten proposed regions: North of ports selected in each region, which has been mentioned
Asia, East Asia, South East Asia, South Asia, Australia, above, Karachi, Mundra and Nhava Sheeva in South Asia
New Zealand and Pacific, Africa, Europe, Middle East, and Aden and Jeddah in Middle East having the highest
North America and Latin America. These regions depend scores are selected.
on the applicable shipping line network. According to the
literature review and interviews with the managers in the
company, the following six criteria are included for the port CONCLUSIONS
selection analysis: location, port charge, port infrastructure,
ship calls, container traffic, and water depth. In this system, the most valuable trait is that the user
himself can select the port criteria he wants and change the
Selection of the port region, the number of ports visited preference of each criterion according to the real situation
and the subject criteria to be analyzed is done by the first and company policy.
interface of the system. Through this interface, the decision
maker selects the port regions to be served in the liner Due to the flexibility of the system, the decision maker
network analysis by adding a new port region or removing in the company can change his preferences case by case,
any port region from the port regions list. After selecting which helps to enhance the service quality of the liner
port regions to be served, the decision maker chooses the shipping company and get a more competitive position in the
number of ports to be visited in the responding port region dynamic shipping market. The paper shows how technology
and the subjective criteria to be considered for the liner advancement like utilising DSS can bring positive effects
network within these selected port regions. on strategic planning of shipping companies.
In this case, the user of the system needs to choose 3 The work to develop the DSS can be enhanced in future
ports in South Asia and 2 ports in Middle East to serve a research. Port criteria are essential to select ports and
liner network. A liner network planning manager (decision represent an important linkage in the DSS. Marketing theory
maker) in the company bases on company policy and/or his has pointed out that a customer’s perception of a particular
experience and selects three criteria – port infrastructure, situation is often different from that of another customer.
port charge and container traffic - to evaluate the ports. Since Therefore, future work includes surveying the perspective
the criteria selection is based on company’s/ management’s of the shippers in various port regions, and modifying the
preference, other users can choose different criteria for port system according to the new port criteria in order to give
evaluation to get different results. the user more precise choices on port criteria.
89
DANGEROUS GOODS REGULATING

SYSTEM IN SINGAPORE
Cui Yifang (cuiy0003@ntu.edu.sg)
Wong Yiik Diew (cydwong@ntu.edu.sg)
ABSTRACT: Dangerous Goods (DG) can lead to serious consequences if improperly managed. Hence, a sound regulating system is
needed to safeguard the handling of DG. In Singapore, many international and national regulations have been implemented. Supportive
initiatives have also been launched by DG agencies to improve the regulation system.
Our research study shows that there are incongruous opinions between industry and regulating agencies concerning Singapore’s DG
regulation system. While industry companies aspire for a unified system with one DG agency in charge to reduce confusions existing in
the current system, the regulatory bodies have their reasons to retain the multi-agency system as they continue to improve and delegate
responsibilities clearly among DG agencies.
A common response from the industry interviewees is that there is room to improve the effectiveness of communication between industry
and regulating agencies. The agencies shall endeavour to timely inform companies about DG developments.
INTRODUCTION and applied technologies to help the industry to build up

a strong DG logistics chain. At the same time, the DG
Dangerous goods (DG) are radioactive, flammable, explosive industry is also paying good attention to the handling
or toxic substances and organisms in solid, liquid or gas process of DG.
forms that can cause danger to the public, property and the
environment (UOW 2007). As a result, there are a number The objectives of this study are to identify critical
of regulations covering safe transportation, storage and issues for DG regulating system in Singapore through a
packaging of DG such as International Maritime Dangerous thorough literature review, and to gauge the usefulness
Goods (IMDG) Code and Maritime and Port Authority of and compliance status of DG regulations by assessing the
Singapore (Dangerous Goods, Petroleum and Explosives) logistics industry’s perceptions on the current regulation
Regulations. system for DG transport in Singapore.
Globally, DG is heavily regulated especially in Europe as The research is within the context of Singapore with focus
it concerns the health and safety of the population as well on transportation and logistics of DG. Relevant regulations
as being environmental hazards. Moreover, DG can also and supporting activities are reviewed and profiled; the
become terrorist weapons which could cause potentially compliance status and usefulness of regulations, as well
dangerous situations with disastrous consequences. as the future trends, are obtained from interviews and
survey.
In Singapore, DG carriers can be seen frequently on the
roads with special signs indicating the specific class of
DG on board. Since there is limited geographical space in METHODOLOGY
Singapore, even approved DG vehicles transportation routes
cannot avoid the closeness to the residential and central Primary data for this study were collected through interviews
district areas. Furthermore, research shows an increase and surveys. Logistics companies in Singapore which are
in frequency of accident occurrences in transportation of involved in DG transportation or handling were contacted.
DG in Europe and North America regions ever since the
Interviews were conducted with DG professionals in the

beginning of 20th century to 2004 (Planas et al. 2008). logistics companies and DG agencies. Survey questionnaires
Singapore is also exposed to similar hazards from frequent were posted to DG logistics companies in Singapore. Survey
DG transportation and storage given its small land space. questions were designed in accordance to the objectives
The hub-and-spoke system that the port of Singapore has of this study. Online survey was also administered on the
adopted requires an enduring, safe and secure transition of survey targets.
cargoes including DG which is a competitive advantage of
Singapore port compared with the others. Therefore, the Secondary data collection covered literature reviews
importance of a well-managed DG system is enormous. of various resources such as databases, journals, books
In order to safeguard the process of transport, storage and and internet. Secondary data are important to generate
carriage of DG, several governmental agencies in Singapore supportive evidence and comparative references for this
have launched initiatives including various regulations research study.
90
LITERATURE AND INTERVIEW FINDINGS Primary activities and supporting activities
Several prior studies were assessed in order to understand Primary activities refer to the international and national
DG regulating system in Singapore, and thus to identify regulations while supporting activities serve to strengthen
the critical issues inherent in the system. implementation of the regulations. The main supporting
activities discussed in this study are technology applications
such as vehicle tracking devices and web portals, while
General system other supportive initiatives are conferences and responsible
care held and promoted by Singapore Chemical Industry
Having adopted a number of international DG rules and Council (SCIC). The characteristics of the primary and
regulations, Singapore has steadily improved its DG secondary activities can be categorised as follows:
regulating system and is catching up on European countries. • General rules and regulations
Mr. Jacobsen from Leschaco Pte Ltd observed, during an
• Classification and labelling of DG
interview, that Singapore has made considerable progress
in the past decade especially in the aspect of warehousing • Declaration of DG
for DG. • Emergency Management
• Licensing Control
In general, Singapore has established a relatively sound • Health and safety of personnel
DG regulating system based on the various regulations
and supportive activities promoted by several DG agencies. These categories of initiatives were surveyed by compliance
These regulations cover different aspects of DG transport status and usefulness, and the results are presented in the
and logistics which safeguard the DG transport chain within following Survey and Interview Results section.
Singapore in air, land and ocean freight. As Mr. Foong from
DHL has mentioned, Singapore enjoys a good international
reputation with good ratings for quality, technology and DG agencies in Singapore
extremely low corruption as well as good transportation
infrastructure. These performance indicators are important
The Singapore’s DG regulation system encompasses
considerations in the management of DG because it gives
a multiplicity of agencies (Table 1) which can cause
confidence to customers that DGs are managed in an
inconvenience for DG handling parties such as manufacturers,
efficient and effective manner with full compliance to
carriers and storage providers. Although each agency has
international DG regulations.
clearly-defined roles and responsibilities, there are grey
areas in the regulating system due to the complexities of
DG properties that result in different standards.
DG incidents in Singapore
Table 1. DG agencies in Singapore.
Internationally, DG accidents occur most frequently in
Main DG agencies
developed countries, especially on highways. The Major
Singapore Civil Defence Force (SCDF)
Accident Hazards Bureau (MAHB) in Europe was
National Environmental Agency (NEA)
established for reporting and analysing DG accidents. In
Singapore Police Force (SPF)
Singapore, as noted by several interviewees, the accident
rate is low compared with other developed countries. Other DG agencies
This may be due to two reasons-one being Singapore’s Ministry of Manpower (MOM)
limited landscape but more importantly, implementation Maritime and Port Authority of Singapore (MPA)
of prevention measures and a well-established regulation Health Science Authority (HAS)
Singapore Customs (SC)
framework such that possibility of accident occurrence is
Land Transport Authority (LTA)
minimised. However, without an established accidents SPRING Singapore
reporting system, accurate accident data are not available
PSA
for analysis. Mr. Tang from DGM Pte Ltd said that the
scale of DG incidents in Singapore was usually small and
Mr. Heng of SCDF ascribed the principal cause of incidents SUMMARY OF SURVEY AND INTERVIEW
to human negligence.
RESULTS
A strong emergency response programme is considered
The perspectives of the industry towards the inconvenience
as critical for dealing with DG incidents. In Singapore,
inherent in multi-agency regulation system were garnered
according to Mr. Kwok of SCDF, there are 4 Hazmat
from the survey. The scope of a compromise between the
stations located at Alexandra, Jurong Island, Tuas and
agencies and the industry towards enhancing the efficiency
Tampines. The Hazmat stations together with fire fighting
of Singapore’s DG regulation system was explored.
stations and Company Emergency Response Team (CERT)
form a strong emergency response team.
91
Survey results continue strengthening the structure, and improving

clarity of each agency’s roles and responsibilities.
Most participating companies in the survey are involved b. Interviewees had different perspectives regarding
in ocean freight, transportation and logistics and almost coverage of regulations. Some perceived grey areas
half of the respondents are handling all 9 classes of DG. that warranted further clarification by the agencies;
The most commonly handled class of DG is class 3 which some felt that there were duplications in coverage of
is flammable liquids. The survey obtained the following regulations as well as responsibilities of the agencies.
observations: Moreover, as Mr. Jacobsen mentioned, in comparison
a. Most respondents agreed that the numbers of with Europe, Singapore still needs to establish more
regulations, regulatory agencies and web portals in regulations, especially certain in-house practices. He
Singapore are adequate. suggested that Singapore should introduce a mandatory
b. There was contention regarding coverage of the appointment of a Dangerous Goods Safety Advisor,
regulations. One in four respondents perceived that the who is trained and certified based on the local and
regulations do not fully cover all applicable DG issues; international rules and regulations, similar to the EU
42% of the respondents considered comprehensive practice.
coverage while the remaining 33% of the respondents c. Interviewees suggested that more educative information
felt that there were duplications in some areas. could be included in web portals for amateur to learn
c. Half of the participants rated the regulations as being about DG on-line. Moreover, vehicle tracking devices
somewhat useful while the other half rated very useful. are only installed on vehicles licensed to carry DG, and
Most participants viewed web portals and the vehicle non-licensed DG carriers continue to pose threats.
tracking devices as being somewhat useful. d. There was general agreement about the lack of
d. Whereas an early study in 2006 indicated deployment effective communication channel in the DG system.
level being lower than perceived usefulness level in The industry may not have timely awareness of new
various categories of activities launched by relevant DG updates in regulations or functions on the web portals,
agencies, the present study revealed comparable levels as a result, some inconvenience would arise.
between deployment and usefulness levels, thereby e. The cost associated with compliance of regulations
suggesting improvements in the activities of the DG and deployment of technologies was considered
system in Singapore over the past few years. relatively high by the interviewees, for example,
e. All the R&D needs proposed to the respondents were the infrastructure cost and training cost. However,
viewed as important and urgent. These R&D needs effectiveness, usefulness and safety considerations
included integrated DG regulation system, regulations seemed more important than cost. Nevertheless, more
for post-accident environmental protection, strong cost-effective measures should be developed by the
accident-reporting system and communication channel agencies for the benefit of the industry.
between regulatory bodies and industry for relevant DG
information updates. In the 2006 study, multiple DG
agencies and weak communication in disseminating CONCLUSIONS AND RECOMMENDATIONS
DG regulation updates were also rated as important
weaknesses existing in Singapore. The major findings from this study on Singapore’s DG
regulating system are summarised as follows.
Interview results First of all, Singapore has an established DG regulation

system encompassing different initiatives and agencies.
The interview results further complement the survey results However, in comparison with the European system, there
from industry and regulatory agencies. is still space for Singapore to improve further. One major
finding is the industry’s claims of multiple DG agencies in
a. Interviewees expressed a multitude of opinions Singapore to regulate the system which lead to confusions
regarding the multi-agency issue in Singapore. and time consumption. The industry aspired for an integrated
Interviewees who considered the number of agencies system to make transactions more effective and efficient
as being too many were principally concerned that such as when applying for various licences. Moreover, this
there would be some repetition in the works to be can also help companies to reduce consultation times with
done such as when applying for licences, submission different agencies when enquiring on DG issues. Contrarily,
of DG manifest, etc. Interviewees who considered the DG agencies stated that they had recognised the benefit of
number of agencies as adequate held the view that such system to industry and had discussions on this issue for
principal DG regulating agencies, namely the SCDF, the past few years. The agencies have decided to maintain
NEA and SPF, should not thus pose any big problem. a multi-agency framework whereby each agency has DG
On the other hand, the agencies did not foresee any professionals to deal with specific kind of DG matters.
major change to the current framework, and they would Furthermore, there are only three major DG agencies namely
92
SCDF, NEA and SPF, and they constantly endeavour to there is less confusion. This intended action would affect
define division of roles among each agency clearly for the the national standards as many countries are complying
benefits of industry. The multi-agency approach involving with both UNRTDG and GHS. Moreover, Singapore port’s
each agency exercising responsibilities for specific DG indigenous classification of DG, i.e. the PSA Classes 1,
issues shall be maintained in the foreseeable future. 2 and 3, which only apply in Singapore port may cause
misunderstanding by shipper and carriers of DG, especially
Secondly, one important and urgent demand from the if they are unfamiliar with the system.
industry is effective communication with the authorities.
Updates on the web portals or regulations may not be This research study has mainly focused on the internal
timely communicated to the industry thereby causing factors, specific to Singapore’s context. Future research
inconvenience and delay. Previously, it was SCIC playing studies on DG regulating system should include more
the bridging role in bringing industry and agency personnel external factors such as comparisons with other countries’
together to share opinions and suggestions several times DG systems. The advantages and disadvantages of each
each year. One key initiative should be to promote and country’s DG framework can be examined to find the gaps.
enhance connection between the industry and the regulatory By analysing both internal and external environments, the
bodies to facilitate the process of exchanging information structure and content of Singapore’s DG regulating system
and points of views. can be further improved to achieve an even safer and more
effective DG regulating system in Singapore.
Thirdly, by comparing the findings of this study with the
2006 study by Mr. Rajkumar on DG logistics system in
Singapore (Rajkuma 2006), there has been better industry REFERENCES
compliance with agencies’ initiatives as resulting from the
continuous efforts put in by both the companies and the [1] Planas, E., Pastor, E., Presutto, F. and Tixier, J., 2008. “Results
agencies to improve DG system in Singapore. of the MITRA project: Monitoring and intervention for the
transportation of dangerous goods”. Journal of Hazardous
As for further R&D, one potential area is the harmonisation Materials, 152(2), 516-526.
of DG classification system at an international level. As [2] Rajkumar, T.V., 2006. Analysis of Dangerous Goods Logistics
there are two sets of chemical classification systems defined in Singapore. Master of Science (Logistics) dissertation,
by United Nations Recommendations on the Transport of Singapore: Nanyang Technological University, retrieved
Dangerous Goods (UNRTDG) and United Nations Globally September 2009.
Harmonised System of Classification and Labelling of [3] University of Wollongong, 2007. Dangerous Goods, School
Chemicals (GHS), international organisations do aspire and of Chemistry, Australia: University of Wollongong, retrieved
have intention to unify the classifications system so that November 2010.
93
DETERMINATION OF COEFFICIENT OF
CONSOLIDATION BY ROWE CELL
Budi Wibawa (cwibawa@ntu.edu.sg)
Liyenita Widjaja (liye0002@ntu.edu.sg)
ABSTRACT: Structures built on clay layers will cause consolidation settlement. There are two aspects of consolidation settlement,
i.e. the magnitude of settlement that is related to compression index and the rate of settlement corresponding to the coefficient of
consolidation. A Rowe cell was used to determine the coefficient of consolidation for both vertical and horizontal flow directions with
various drainage conditions.
INTRODUCTION the need of the determination of both coefficients of

consolidation, cv and ch. While cv is usually determined by a
Structures which are built over soil layers will experience conventional oedometer test, ch cannot be done by this test;
settlements. One of the settlements, consolidation settlement, therefore a Rowe Consolidation Cell (Rowe and Barden,
will mainly occur if the structures rest on saturated clay 1966) was used instead of the oedometer, to determine the
layers. Consolidation settlement is related to consolidation coefficient of consolidation. Hence, the objective of this
- a process of reduction of volume due to dissipation of pore- article is to determine the coefficients of consolidation of
water pressure in the clay layers due to induced compression. clay in both vertical and horizontal directions using the
There are two aspects of the consolidation settlement - the Rowe cell for various drainage conditions.
magnitude and rate of settlement. The magnitude of the
consolidation settlement is related to compression index,
whereas the rate of settlement corresponds to coefficient of EXPERIMENTS AND DISCUSSION
consolidation. The coefficient of consolidation is a measure
of the rate of consolidation or the rate of water flow inside Material
the soil, which is used to determine consolidation time. In
this article, our focus will be on the determination of the Kaolin was used for the experiments. Tests on basic
coefficient of consolidation of saturated clay. properties of soil were done in accordance to BS: 1377
– Part 2. The soil properties are Specific gravity (2.65),
Terzaghi et al (1967) indicated that both permeability (k) Liquid Limit (76.5%) and Plasticity Index (27.3%).
and coefficient of compressibility (mv) decrease rapidly Three tests to determine coefficient of consolidation were
with decreasing void ratio so that the ratio (k/mv) tends conducted, i.e. single drainage, double drainage and radial
to be constant; hence, the coefficient of consolidation in drainage conditions.
vertical flow direction (cv) is fairly constant over a wide
range of effective vertical consolidation pressures. However,
based on a study conducted by Robinson & Allam (1998) Single drainage
on the effect of clay mineralogy on the coefficient of
consolidation, it was found that cv is not constant, but The values of cv are found to be in the range of 8.05×10-7 to
varies with consolidation pressures. It was also found 4.33×10-6 m2/s. Although Terzaghi assumed the permeability
that the compressibility of clays is influenced by both (k) and the coefficient of volume compressibility (mv) to
mechanical and physicochemical effects (Mesri & Olson, be constant, during consolidation process, as water is being
1970), depending on the type of mineral, saturating cation squeezed out from the soil, the permeability of the soil
and the pore fluid. Sridharan et al. (2004) observed that soil will decrease, as well as the coefficient of compressibility.
with lesser shrinkage index or plasticity index shows an However, the rate of decrease of permeability and
increasing trend of cv versus vertical effective stress (σv’), compressibility may not be the same (Robinson & Allam,
but for higher shrinkage index or plasticity index, there is 1998), and the difference in the rate affects the trend of
a decreasing trend of cv versus σv’ for soils of nearly the cv. In this experiment, the trend of cv tends to increase in
same liquid limit. the early stage of the consolidation up to a certain value,
and then decrease. This may be due to a larger rate of
In a saturated clay layer, water may flow in both horizontal decrease of permeability as compared to the decrease of
and vertical directions during consolidation process. The compressibility. When this trend occurs, even if there is a
rate of flow of water in vertical and horizontal directions decrease in both permeability and compressibility, the ratio
contributes to the overall rate of consolidation indicating of the rate may give an increase to the cv.
94
Double drainage CONCLUSIONS
For double drainage tests, the trend of cv is found to have In this study, Rowe consolidation cell was used to overcome
nearly the same trend as that of single drainage, but it has the major disadvantages of conventional oedometer test.
a slightly lower value. The values of cv are in the range of Rowe cell allows a larger sample size, which is very
5.19×10-7 to 1.65×10-6 m2/s. During consolidation process, advantageous in the case of testing a non-uniform soil
both permeability and compressibility reduce with increasing sample. On top of that, the application of load is simpler,
effective consolidation pressure. The reason for this trend and it allows a control of drainage.
may be due to the significant difference between the rates
of decrease of permeability as compared to the rate of The coefficient of consolidation in the vertical flow direction
decrease of the coefficient of compressibility. In the early range is to be from 8.05×10-7 to 4.33×10-6 m2/s in the single
stage of consolidation, the rate where permeability decreases drainage condition and from 5.09×10-7 to 1.65×10-6 m2/s
is larger than the rate of decrease of the compressibility. in the double drainage condition.
When the rate is almost the same, the ratio is constant and
it will cause the cv to decrease with increasing effective In addition, the coefficient of consolidation in the horizontal
pressure. The soil properties as compared to the specimen flow direction is found to be from 4.29×10-7 to 7.85×10-7
of Sridharan (2004) are similar. As the soil specimen has a m2/s in the radial drainage condition.
higher liquidity index (IL) when cv is directly proportional
to IL, consequently, the calculated values of cv in this
experiment are higher than that of Sridharan (2004). REFERENCES
[1] Chu, J., Myint, W.B., Chang, M.F. and Choa, V., 2002.
Radial drainage “Consolidation and permeability properties of Singapore
marine clay”. Journal of Geotechnical and Geoenvironmental
Engineering, Vol. 128, No. 9, pp. 724-732.
The coefficient of consolidation in the horizontal flow
direction (ch) has a range between 4.29×10-7 and 7.85×10-7 [2] Mesri, G. and Olson, R., 1970. “Mechanisms Controlling the
Compressibility of Clays”. Journal of American Society of
m2/s. It was found that ch shows a decreasing trend with
Civil Engineers, Vol. 96, pp. 1853-1878.
increasing effective consolidation pressure, and generally
lower value than that of cv. The decrease is noted to be more [3] Robinson, R. and Allam, M., 1998. “Effect of clay mineralogy
on coefficient of consolidation”. Clays and Clay Minerals,
visible in low pressure range, whereas in high pressure, it
Vol. 46, No. 5, pp. 596-600.
tends to increase. As compared to the research on Singapore
[4] Rowe, P. and Barden, L., 1966. “A new consolidation cell”.
marine clay done by Chu et al. (2002), the values of ch of
Géotechnique, Vol. 16, No. 2, pp. 162-169.
kaolin in this study are found to be higher than Singapore
[5] Sridharan, A. and Nagaraj, H., 2004. “Coefficient of
marine clay. However, the trend of ch with respect to
consolidation and its correlation with index properties of
effective consolidation pressure is in a good agreement with remolded soils”. Geotechnical Testing Journal, Vol. 27, No.
previous findings. As the effective consolidation pressure 5, pp. 469-474.
increases, both permeability and compressibility will also
[6] Terzaghi, K., Peck, R.B. and Mesri, G., 1996. Soil Mechanics
decrease. In this case, it is probable that the rate of decrease in Engineering Practice. New York: Jon Wiley & Sons,
of permeability and compressibility are relatively the same, Inc.
which leads to a decrease in coefficient of consolidation.
The different value may be due to the difference in the
nature of the soil.
95
EFFECTS OF ELECTRIC VEHICLES ON

CLIMATE GOALS – SINGAPORE AND
GERMANY IN COMPARISON
Rainer WITZIG (rainer.witzig@tum.de)
WONG Yiik Diew (cydwong@ntu.edu.sg)
CHANG Wei-Chung Victor (WCChang@ntu.edu.sg)
ABSTRACT: We create a model to simulate the effects of a large number of electric vehicles (EVs) on energy demand and CO2-emissions.
The model reflects the impacts on electricity demand, the structure of power plants, electricity price, and the price elasticity of cars. We
compared the reduction potential of energy demand and emission in Germany and Singapore.
The simulations show a greater potential to reduce energy demand by using EVs in Singapore than in Germany. The main reason for
this is the higher specific energy demand of conventional cars for solely urban use in Singapore. However, introducing emission-free
power plants in Singapore could improve the CO2-balance further.
INTRODUCTION the changing mileage. The model allows simulations up

to the year 2020. A basic scheme of the EVICE model is
The intensifying climate change comes along with the shown in Figure 1.
need to reduce greenhouse gas emissions significantly. The
transport sector’s substantial role in this context leads to the Firstly, we estimate the mileage of private car traffic in
imperative to decrease the energy demand in transportation Singapore and Germany for the next 10 years. In different
considerably. The changeover from cars with combustion scenarios we assume different numbers of EV entering the
engines to electric vehicles (EVs) is currently strongly market. These substitute conventional cars with a given
supported by politics in many countries, and justified with share of urban /interurban use. Based on these parameters,
the positive effects on climate goals that would come with we compute the electricity demand which is necessary to
it (GEA 2009). This raises high expectations of EVs’ green meet the energy demand for EVs. The model also computes
benefits in media, industries, societies and lobby groups. We the costs to produce electricity, which is one of the two
analysed EVs’ actual impact on climate goals, as well as main parameters to calculate the operating costs of EVs
its side effects, comparing different scenarios in Singapore (Roth 2009).
and Germany.
Since the travel costs driving an EV will differ significantly
to travel costs with a conventional car, the driving
THE EVICE MODEL performance will alter. We model this effect using a given
price elasticity for fuel price. The new mileage EVs results
An argument for the proliferation of electric vehicles in in an electricity demand different to our first assumption.
replacement of fossil fuel-driven vehicles is the comparably Feeding in the new data of electricity demand, we start an
low energy demand for the end user and the theoretical iteration of the steps as shown in Figure 1. After obtaining
option to operate EVs independently of fossil fuels. Also, equilibrium we are able to compute final results for CO2-
due to the current electricity price, operational costs driving emissions, energy demand, mileage and travel cost for EVs
and conventional cars, respectively.
an EV would be much cheaper compared to a conventional

car. However, this could also result in an increasing vehicle
mileage. The future structure of power plants which would
be necessary to meet the increasing demand of electricity SCENARIOS
is another important variable to be analysed. In turn, these
factors would counterbalance today’s linear estimations In a reference scenario, we assume no EV in Germany or
regarding CO2-emissions, electricity price and demand of Singapore in the next 10 years. This allows us to compare
fossil fuels. the results of different settings. In different scenarios we
simulate a substitution of 10% or 100% of conventional
We developed the EVICE model (electric vehicles’ cars by EVs. Whereas the first case seems to be realistic,
impact on climate change and energy demand) for further the last scenario is not lifelike. However, this scenario helps
inspections. This model contains the feedback of electricity to illustrate the maximum potential of EV’s use.
demand, electricity price, price elasticity of mobility and
96
Figure 1. Scheme of the EVICE model.
One important input parameter is the fuel price elasticity mileage. In the scenario with 100% replacement of cars, it
for car use. We assume this value to be -0.05, according would increase by 9% - 27%, depending on the fuel price
to over 100 empirical estimations (Forschung-Informations- elasticity. However, in the more realistic scenario with a
System 2009). 10%-replacement, the increase in mileage is between 1%
and 7% in both countries (see Figure 2).
We also assume a yearly increase of gasoline price by
6% according to estimations of the German Institute 13.5
Singapore
for Economic Research (Hunsicker andSommer 2008). 13.0
However, we also simulate scenarios with different rates and

mileage [billion vkm]
12.5
12.0
elasticities. Further parameters taken into account include
0%
11.5 10%
today’s and future battery costs, average fuel consumption 11.0 100%
10.5
of Germany’s car fleet, share of EVs’ urban/interurban use, 10.0
timeslots for charging, durability of batteries, as well as 9.5

2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
additional taxes for fuel or electricity. The importance of year
these parameters will be discussed in a following sensitivity Germany

analysis. 700
680
mileage [billion vkm]
660
640 0%
RESULTS AND ANALYSIS 620 10%
600 100%
In different scenarios we simulate the impact of a 580
substitution of 10% and 100% of conventional cars by

560
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
EVs in Singapore and Germany. Germany’s composition of year
power plants today consists of nearly 20% renewable energy

Figure 2. Mileage in Singapore and Germany.
resources, which makes the use of EVs environmentally
friendlier. In Singapore on the other hand, the main benefit
of EVs is the fact that the average fuel consumption of The potential savings in energy demand in the 10%-scenario
conventional cars is significantly higher than in Germany. are marginal in Germany. In Singapore on the other side,
The reason for this disparity is not the quality of the car where the average fuel consumption of cars is higher, the
fleet, but the solely urban use of cars, which comes along savings would be 6%, and even 57% in case of a total
with higher specific energy demand. substitution by EVs. In Germany, the according values are
0.5% and 32% (see Figure 3).
First results of the EVICE simulation show that in both
countries the low electricity price would lead to increasing
97
SENSITIVITIES OF UNKNOWN
PARAMETERS
EVICE uses many input parameters which are afflicted with

uncertainty. These include the future fuel price, battery costs,
fuel consumption of the car fleet, timeslots for charging,
restrictions of CO2-emissions for Germany’s power plants,
durability of batteries, as well as additional taxes for fuel
or electricity. We want to analyze the importance of these
variables to the final result. Therefore we run several
simulations, consecutively changing single parameters and
calculate each parameter’s sensitivity. In these simulations,
the reference scenario will be the 10%-replacement of cars
by EVs. The final result which we take for comparison is
the primary energy demand in private transport in 2020.
We compute the sensitivities η using the mean arc elasticity
function:
Q2 – Q1
Figure 3. Primary Energy Demand for motorized vehicles (Q2 + Q1)/2
in Singapore and Germany. ηQ,P =
P2 – P1
(P2 + P1)/2
The CO2 emissions of EV depend on the type of power
ηQ,P : elasticity of changing parameter
plant which is used to generate the electricity. The
Q1 : primary energy demand in10%-scenario
environmental friendly mix in Germany cannot come up
Q2 : primary energy demand with parameter P2
for a 100% replacement of cars by EVs. In consequence,
P1 : parameter in 10%-scenario
new coal fuelled power plants would have to be built. This
P2 : new parameter
shortcoming limits the maximum potential of emission
savings to 37%. In the Singapore case, the CO2 savings
We observe a high relevance of the annual rise in fuel
potential behaves almost linear to the energy demand – its
price, fuel consumption of the conventional car fleet and
maximum potential is around 62%. This can be achieved by
the fuel price elasticity. The other parameters mentioned
continuously using oil and gas fuelled plants. Introducing
above play an insignificant role in the overall outcome.
emission free power plants, this value could even be higher
(see Figure 4).
CONCLUSIONS
We created the EVICE model in order to simulate the

change of all relevant variables when using a large number
of EVs in Singapore and Germany. This includes electricity
demand, the structure of power plants, electricity price, price
elasticity of car use and others. In iterations we calculated
the change of each parameter and its impact on the others,
until equilibrium is reached.
The simulations show a greater potential to reduce energy

demand by using EVs in Singapore than in Germany.
The main reason for this disparity is the higher specific
energy demand of conventional cars in the solely urban

use in Singapore. We also observe a greater potential in
Singapore to reduce CO2-emissions compared to Germany.
The environmental friendly power plants in Germany have
reached the limits of their capacities since not many more
Figure 4. CO2-emission of motorized vehicles wind, water and solar power plants can be built in the future.
in Singapore and Germany. In Singapore, on the other hand, introducing emission free
power plants could improve the CO2-balance even more.
98
The sensitivity analysis showed that the parameters with REFERENCES

the biggest impact in this model are the annual rise in fuel
price, the fuel consumption of the conventional car fleet [1] Forschung-Informations-System. Elastizitäten der
and the fuel price elasticity. Further research is needed to Kraftstoffpreise. http://www.forschungsinformationssystem.
understand the exact reaction of passengers to fuel price de/. 03.06.2009
changes. [2] GEA (2009). Electric Vehicles and CO2 Emissions.
[3] Hunsicker, F. and Sommer, C. (2008). “Welche Zukunft darf’s
denn sein?” Internationales Verkehrswesen Nr. 9, 334-335.
[4] Roth, H. (2009). “Optimierung der Kraftwerkparkentwicklung”.
Lehrstuhl für Energiewirtschaft und Anwendungstechnik, TU
München.
99
MARITIME STUDIES DEGREE

PROGRAMMES IN SHIPPING
MANAGEMENT --
AN INTERNATIONAL COMPARISON
Gou Xueni (goux0001@ntu.edu.sg)
Wong Yiik Diew (cydwong@ntu.edu.sg)
ABSTRACT: World GDP is increasing and international seaborne trade is expanding momentously. Shipping, shipping education and
training have evolved into global and sophisticated businesses. Several maritime degree programmes are offered by universities around
the world for training shipping management personnel who are knowledgeable in both vessel operations and shore-side business. This
study reviews 22 relevant shipping management bachelor degree programmes around the world. The interactive factors include maritime
heritage, economic development, seaborne trade and maritime business. The educational structures and overall academic disciplines of
an institution also have impact on the academic content of these programmes.
INTERNATIONAL ECONOMIC GROWTH MARITIME BUSINESS AND SHIPPING

AND MARITIME BUSINESS MANAGEMENT DEGREE PROGRAMMES
Trade and shipping businesses have always been important Many academies have existed for a long time in training
enablers of wealth and growth. International trade has mariners. As the maritime business evolves, some of
expanded by more than 1,700% over the second half of the these mariners make career transitions from the sea to
20th century in volume terms (WTO, 2010). Although world the shore to helm shore-side shipping businesses. In
GDP experienced a steady growth over time, international light of highly integrated shipping industry, a number of
merchandise trade has roughly tripled in importance academic programmes have flourished to train shipping
compared with the economy as a whole (The World Bank, management personnel. This is evidenced by the existence
2010). It is generally accepted that more than 90% of global and development of maritime degree programmes in
trade is carried by sea. Seaborne trade worldwide has been shipping management in various universities and institutes
growing very rapidly in recent 55 years from 500 million throughout the world.
tonnes in 1950 to 7 billion tonnes in 2005 (UN, 2008).
In order to have a detailed understanding of the shipping
Maritime trade is dominated by three economic centres, management programmes, a sample of 22 programmes is
viz Europe, North America and Asia, strung out along a selected for a substantial analysis, taking into consideration
“Westline”. It is the line along which the commercial centre a combination of factors such as the national economies,
of maritime trade has moved west over the last century maritime heritage, and institutions’ academic disciplines
of maritime business. The recent fast growth in seaborne and curriculum structures.
trade stems especially from Asian countries. Although
European countries keep a large share of seaborne trade,
the growth rate is relatively slow and sea trade volume ANALYSIS
has remained static.
Shipping management programmes and maritime

In terms of fleet tonnage, based on total deadweight ton tradition, economic growth, merchandise trade
(DWT) controlled as a percentage of world fleet by parent
companies located in different countries and territories, the Tradition plays a major part in most European maritime
top 35 countries control 94.0% and 95.6% of world fleet nations. Germany stands out as it experiences fast GDP
as end of 1998 and 2008, respectively. The ship owning growth over the last decade and its merchandise trade
business tends to converge within a small number of as a percentage of GDP is also increasing. Additionally,
countries, among which Asian countries take an increasing based on World Bank Logistics Performance Index 2009,
share from 38.6% to 44.8% (UNCTAD, 2009). Germany has replaced Singapore ranking as world No. 1
(The World Bank LPI, 2009).
100
Table 1. Institutions and degree programmes selected for analysis.

No. Country/Institute/Degree Programme No. Country/Institute/Degree Programme
1 Greece/ 12 Philippines/
University of Piraeus/ Asian Institute of Maritime Studies/
Maritime Studies (4 years) Maritime Business (4 years)
2 Cyprus/ 13 Taiwan/
Frederick Institute of Technology/ National Taiwan Ocean University/
Maritime Studies (4 years) Shipping & Transportation Management (4 years)
3 UK/ 14 PRC/
University of Plymouth/ Shanghai Maritime University/
Maritime Business & Logistics (3 years) Shipping Management (4 years)
4 Germany/ 15 Thailand/
University of Applied Science OOW/ Burapha University/
Maritime Economics & Port Management (4 years) Maritime Management
5 UK/ 16 Japan/
Southampton Solent University/ Kobe University/
Maritime Business (3 years) Maritime Logistics
6 UK/ 17 PRC-HK/
Liverpool John Moores University/ Hong Kong Polytechnic University/
Maritime Business & Management (3 years) International Shipping & Transport Logistics (3 years)
7 USA/ 18 Taiwan/
State University of New York/ Kainan University/
Maritime Studies (4 years) Logistics & Shipping Management (4 years)
8 USA/ 19 Singapore/
California Maritime Academy/ Nanyang Technological University/
International Business & Logistics (4 years) Maritime Studies (4 years)
9 USA/ 20 Singapore/
Massachusetts Maritime Academy/ Nanyang Technological University/
International Maritime Business (4 years) Maritime Studies with Business Major (4 years)
10 USA/ 21 Australia/
Texas A&M University at Galveston/ University of Tasmania/
Maritime Administration (4 years) Maritime & Logistics Management (3 years)
11 Canada/ 22 Egypt/
Memorial University of Newfoundland/ Arab Academy of Science, Technology & Maritime
Maritime Studies (diploma+1 year) Transport/
Management of Trade Logistics & International
Transport (4 years)
Among all the countries where the selected programmes Economic independence after WWII has necessitated the
are offered, the US has the highest GDP but the lowest rapid advancement and expansion of the sea transportation
merchandise trade as a percentage of GDP. Despite the industry in Japan. It ranks No. 1 as of end of 2008
small percentage, the absolute amount of merchandise trade, controlling the largest merchant fleet in the world. Taiwan,
or more specifically seaborne trade, is not little at all. Of Hong Kong and Singapore are Newly Industrialising
US foreign trade, waterborne trade as a percentage of all Economies (NIE) in Asia. They are all surrounded by
modes trade has increased steadily from 2003 to 2008 in sea and control a large fleet by DWT. For Hong Kong
value terms. Domestic waterborne trade is comparable to its and Singapore, there is a huge amount of merchandise
international trade (U.S. Department of Transportation). As trade as a percentage of GDP. Because of their strategic
perceived by Texas A&M University at Galveston, “activity positions, they are striving to become the regional as well as
in American ports is expected to more than double in the international maritime centre. China has indeed a very long
next 20 years, and some ports in the Gulf of Mexico are maritime history. Compounded by its rapid growth in GDP
expected to see a tripling of port activity”. In addition, and international merchandise trade since joining WTO in
the next 40 years will see greatly expanded oil and gas 2001, China is emerging as a major maritime nation. Over
production in the deep waters of the western Gulf of Mexico 90% of its foreign trade volume is transported by sea and
(Texas A&M University at Galveston). shipping on sea and on river also plays an important role
for domestic traffic. Its merchant fleet also ranks among
Japan outstands in Asia with very high GDP. It gained one of the world’s largest five.
maritime power at almost the same time as the West.
101
Course categorisation Analysis in groups
The courses within each programme are categorised The selected programmes are classified into three groups
into maritime-technology, business-management, hybrid (Figure 1). Group A (No. 8-10, 12-14, 22) includes
and general. Maritime Technology (M-T) courses are programmes which are offered in independent maritime
courses which are specific to those in shipping and institutes, Group B (No. 5-7, 11, 15, 21) consists of
would not normally be offered in other programmes, e.g. programmes that are offered by maritime colleges or
Ship Technology, Maritime Transportation, etc. Business academies within comprehensive universities, and Group C
Management (B-M) courses are courses that are typically (No. 1-4, 16-20) comprises programmes in comprehensive
found in business and management programmes, e.g. universities which do not have any specific maritime
Accounting, Economics, etc. Hybrid courses are those college or academy (with one exception: No. 1 is offered
courses that have elements of both M-T and B-M contents, by a university as a principally business discipline). As
e.g. Shipping Economics, Maritime Law, etc. General an overview, B-M and Hybrid courses are the dominant
courses refer to courses that are not specific to any degree course types.
programmes whereby the basic premise is for education
broadening, e.g. Physics, Chemistry, etc. All the figures B-M courses dominate in Group A; while B-M and Hybrid
are expressed percentagewise below. The academic units courses dominate in Groups B and C. Group C has more
for each course are directly available according to the M-T courses as compared with the other two groups. The
course content descriptions. Academic units for each type universities/institutes without much maritime background
of courses are then summed and shown as a percentage may place more focus on shipping knowledge as compared
of the total academic units required for the corresponding to the traditional maritime academies. While, with plenty of
programme, as illustrated in Figure 1. expertise in the technical aspects of shipping, it seems more
meaningful for the maritime academies or colleges to put
in more efforts on the business areas. For the independent
maritime institutes (Group A), even the Hybrid category is
emphasized to a lesser extent.
A notable feature of Group A is that General courses have

taken a fairly large weightage. These include 3 programmes
(No. 8-10) selected from the USA and 1 each from China
and Philippines. This is mainly due to the educational
systems in the respective countries which give much focus
on the general tertiary education. This is evidenced by
No. 7 programme under Group B which is offered in the
USA by SUNY Maritime College, part of a comprehensive
university now, and it has heavy General course content as
well. Meanwhile, this programme has clearly more General
courses and Hybrid courses but fewer B-M courses when
compared with the other 3 USA programmes. Drawing from
this case, it is perceived that the composite and organisation
of the education systems certainly have some impact on
the structures of various shipping programmes.
As for the extraordinariness of programme No. 16, it is

worthwhile to point out more background information.
Although Kobe University is considered as a comprehensive
university with no specific maritime college or academy, the
Faculty of Maritime Science (established in 2003) indeed

can be traced back to 1917 with its origin as Kawasaki
Merchant Marine School. It was nationalised in 1920 as
Kobe Nautical College and later united with other two
nautical colleges in Tokyo and Shimizu to become as
Kobe National College (Kobe University). In this sense,
the Kobe University should have plenty of expertise and
resources in M-T education which they exploit to the extent
that they outstand in M-T courses weightage among other
comprehensive universities.
Figure 1. Course content by groups.
To further gauge the breadth and depth of maritime and
business knowledge in different shipping management
102
Figure 2. Comparison of M-T content and B-M content.
programmes, comparison is made on the basis of a REFERENCES

conceptual framework as developed by Leong, Wong &
Williams (2009). Academic units for Hybrid courses are [1] Kobe University. Retrieved on 12 May, 2010 from http://www.
divided equally between the M-T and B-M categories maritime.kobe-u.ac.jp/maritime_e/history.html
while general courses are excluded in this comparison. [2] Leong, E.C., Wong, Y.D. and Williams, E.C., 2009. Conceptual
The percentages are calculated and shown in Figure 2. Framework for Comparing University Baccalaureate
Generally speaking, B-M content is being given more Programmes in Shipping Management. WMU Journal of
attention than M-T content. Maritime Affairs 8(1): 47-58.
[3] Texas A&M University at Galveston. Retrieved on 31 May,
2010 from http://www.tamug.edu/mara/academic%20 program.
htm
CONCLUSIONS
[4] The World Bank Statistics, 2010. Retrieved on 26 June, 2010
from http://databank.worldbank.org/ddp/home.do?Step=12
There are strong interdependent links between economic
andid= 4&CNO=2
growth, global trade, seaborne trade and maritime transport
[5] UNCTAD Review of Maritime Transport, 2009. Retrieved
service. With transformation of the management structures
on 26 June, 2010 from www.unctad.org/en/docs/rmt2009_
in the shipping industry, the education system for training en.pdf
shipping management personnel has evolved and will
[6] United Nations Statistical Yearbook, 2008. 52nd Issue. United
continue to evolve. The westerly shift of maritime power
Nations: New York.
has prompted the establishment of various shipping
[7] U.S. Department of Transportation Maritime Administration,
management programmes in Europe, North America, Asia
2009. U.S. Water Transportation Statistical Snapshot. Retrieved

and elsewhere. They are offered in maritime academies as on 20 May, 2010 from http://www.marad. dot.gov/documents/
well as universities and institutes. A substantial sample of US_Water_Transportation_Statistical_snapshot.pdf
22 shipping management bachelor degree programmes is [8] The World Bank Logistics Performance Index, 2009. Retrieved
selected for an international comparison. The interactive on 26 June, 2010 from http://info.worldbank.org/etools/
factors include maritime heritage, economic development, tradesurvey/mode1b.asp#ranking
seaborne trade and maritime business. Business content [9] WTO Statistics, 2010. Retrieved on 26 June, 2010 from http://
is paid greater attention, and Hybrid courses provide an www.wto.org/english/res_e/statis_e/its2010_e/its2010_e.pdf
integrated means to bridge the general business and maritime
business. In essence, the academic content of the different
shipping management programmes varies depending on
the respective educational structures and overall academic
disciplines of an institution.
103
P-WAVE VELOCITY MEASUREMENTS

IN SEDIMENTARY ROCKS
Wong Ngai Yuen Louis (LNYWong@ntu.edu.sg)
Zhang Xiaoping (ZhangX.P@ntu.edu.sg)
ABSTRACT: We investigated the P-wave velocity of sedimentary rocks, including sandstone and siltstone, in natural water content and
saturated conditions. The P-wave velocity measurements indicated that the anisotropic characteristics of siltstone were more significant
than that of sandstone. The correlation between P-wave velocity and uniaxial compressive strength (UCS) was also studied.
INTRODUCTION Except in the immediate vicinity of the seismic source, i.e.

transmitter, the strains associated with the passage of seismic
As part of the on-going experimental studies to characterize pulse are minute and may be assumed to be elastic. Based
the mechanical behavior of the sedimentary rocks in on this assumption, the propagation velocities of seismic
Singapore, P-wave velocity tests were conducted. This pulse are determined by the elastic moduli and densities
non-destructive geophysical testing method offers a reliable of the materials through which they pass. The velocity of
means to reveal the anisotropic nature in the rocks. It propagation of a compressional body wave, i.e. P-wave
also has a promising potential to serve as a quick means (Vp) in any material is given by
to correlate the strength of the rocks with the measured
E
P-wave velocity. Vp = √ p …(1)
Two major rock types, siltstone and sandstone, were tested. where E is the Young’s modulus and ρ is density of material.
To help reveal rock anisotropy characteristics, P-wave The velocity of a shear body wave (Vs), which involves a
velocity tests were conducted in a direction generally pure shear strain, is given by
parallel to the sedimentary plane (θ=0°) and perpendicular
to the sedimentary plane (θ=90°). The definition of θ is Vp = √ G
p …(2)
shown in Fig. 1.
Receiver where G is the shear modulus. Since the first transmitted

Transmiter Specimen arrival wave is the P-wave, its detection is relatively easy.
The shear-wave arrival, however, may be obscured by
© vibrations due to ringing of the transducers and reflections
Se of the compression wave. Therefore P-wave (compression
dim
ent wave) is more favorably used in laboratory rock testing.
ar y
p la
ne A number of studies have found that P-wave velocity and
rock properties are closely related. From testing of nineteen
Figure. 1 Schematic illustration of P-wave velocity test. different rock types, Kýlýç and Teymen (2008) obtained
the correlation between P-wave velocity and mechanical
METHODOLOGY properties, which includes UCS (Fig. 2a), indirect tensile
strength (Fig. 2b) and loss of volume (Fig. 2c). The latter
Principle is a measure of abrasion resistance. Kahraman and Yeken
(2008) also identified strong correlations between P-wave

Ultrasonic techniques are widely used in geotechnical and velocity and physical properties of rock, which includes
rock mechanics applications as they are non-destructive density, porosity, void ratio, water absorption by weight
and easy to apply in both site and laboratory conditions. based on fourteen different carbonate rocks.
The sound velocity of a rock mass is closely related to
the intact rock properties. Measuring the velocity in rock
masses can interrogate the rock structure and texture. The Methodology
important influencing factors are rock type, mineralogical
composition, rock texture and structure, grain size and The instrument used for the present P-wave velocity
shape, density, porosity, anisotropy, porewater, weathering measurement was the CNS Farnell Pundit Plus (Model
and alteration, bedding planes, joint properties (roughness, PC1006), as shown in Figure 3. CNS Farness ultrasound
filling material, water, dip and strike, etc.), in addition to couplant was applied at the end faces to facilitate the
confining pressure and temperature. coupling effect between the rock core and the sensors.
104
Figure 3. Experimental set-up for P-wave

velocity measurement.
(a)
measurements were conducted in 92 specimens, in which

47 of them can be classified as siltstone (or siltstone with
occasional beds of sandstone), while 45 of them can be
classified as sandstone. The specimens are either of natural
water content (considered to be in dry condition) or in wet
condition. The wet specimens had been immersed in water
under a vacuum condition for 26 days before the P-wave
velocity measurement was performed.

(b)
P-wave velocity measurements
The test results of the P-wave velocity measurements

are summarized in Table 1. In general, the average P-
wave velocity values in sandstone are higher than those
in siltstone, in orientation both along and normal to the
sedimentary planes.
For siltstone of natural water content, along the sedimentary

plane (θ=0°), the average velocity is 5679.3 m/s, which
is higher than that of normal condition (θ=90°) – 5405.6
m/s. For sandstone of natural water content, the velocities
in both directions are close to each other, 5838.5 m/s and
(c) 5829.9 m/s respectively. The relatively larger difference
of P-wave velocity in the two measured orientations of
Figure 2. Empirical relationships between direct P-wave
velocity and (a) uniaxial compressive strength (UCS),
siltstone compared to that in sandstone may be attributed
(b) indirect tensile strength and (c) loss of volume (Kýlýç to the stronger degree of anisotropy in siltstone than
and Teymen, 2008). in sandstone. In soaked condition, the average P-wave
velocities of siltstone & sandstone in orientation both along
and perpendicular to the sedimentary plane all increased
The testing procedures were in general accordance with
as compared to the dry condition. The Vps/Vpd ratio is also

the first method contained in the “Suggested Methods computed for reference (Table 1).
for Determining Sound Velocity” of Ulusay and Hudson
(2007).
Correlation with rock strength
The P-wave velocity was calculated from the following
equation: Correlations of the P-wave velocity with the strength values
of the rock specimen were also attempted in the present
Vp = d/tp …(3) study. The first correlation was based on the empirical
relationships of equations (4) and (5) provided by Kýlýç
where Vp is the P-wave velocity, d is the length of the and Teymen (2008), as shown in Fig. 2, for the prediction
specimen and tp is the time taken by the P-wave to travel of the uniaxial compression strength (σc) and the tensile
across the specimen from transmitter to receiver. The strength (σt). The results are listed in Table 2.
105
Table 1. P-wave velocities in siltstone and sandstone.

P-wave velocities (m/s)
θ Natural water content (Vpd) Soaked (Vps) Vps / Vpd
Maximum 5687.6 5520.9
90° Minimum 5126.0 5259.6
Average 5405.6 (16) 5426.0 (12) 1.004
Siltstone
Maximum 5869.0 5986.2
0° Minimum 5571.0 5575.7
Average 5679.3 (11) 5732.8 (8) 1.009
Maximum 6030.5 5878.5
90° Minimum 5585.6 5796.1
Average 5829.9 (16) 5845.6 (12) 1.003
Sandstone
Maximum 5995.8 6070.9
0° Minimum 5611.2 5757.1
Average 5838.5 (9) 5921.8 (8) 1.014
Note: the number in parenthesis after each average velocity value indicates the number of specimens
measured in each group.
σc = 2.304Vp2.4315 (R2=0.94) …(4) Table 3. Test results of P-wave velocity (Ma et al., 2010).
Specimen Saturated Dry condition
σt = 0.49 Vp1.8723 (R2=0.92) …(5) number condition
1 5000 4761
The average UCS values of siltstone obtained from
correlation are 139.4 MPa and 157.2 MPa for θ=90o and 0o 2 5263 5000
respectively. For sandstone, the correlated UCS values are Siltstone 3 5263 5000
167.6 MPa and 168.2 MPa for θ=90o and 0o respectively.
4 5000 4761
For both siltstone and sandstone, the UCS values obtained
based on the correlation appear to be lower than those 5 5882 5555
obtained from the previous UCS experiments performed Ave 5282 5015
on similar rock types (Ma et al., 2010). A major reason
1 5555 5263
is that the empirical relations, which serve as a basis for
the above prediction (Table 2), is derived from a data set 2 5263 5000
consisting of nineteen different rock types, instead of the Sandstone 3 5882 5555
sedimentary rock types specifically tested in this project.
4 5263 5000
Table 2. Estimated UCS and σt based on average 5 5555 5263
P-wave velocity measurements. Ave 5504 5216
P-wave Estimated Estimated
θ
velocity (m/s) UCS (MPa) σt (MPa) specimens in this batch, three specimens, A3, A4 and A5
90° 5405.6 139.4 11.5 tested in the previous phase were included in the present
analysis.
Siltstone 0° 5679.3 157.2 12.7
90° 5829.9 167.6 13.3 The UCS (MPa) data and the P-wave velocity (m/s) data
are plotted in Figures 4 and 5. Due to the substantial degree
Sandstone 0° 5838.6 168.2 13.3 of scattering, and most importantly very small sample size
(only 5 in each group), a statistically significant correlation
A second approach of correlation, which was based on cannot be derived with confidence. To obtain an acceptable
the experimental results previously obtained by NTU (Ma correlation the sample size has to be increased, which will
et al., 2010), was also attempted. In this previous study, be the scope of work in the near future.
twenty specimens equally divided into four groups were
tested, which included dry siltstone, saturated siltstone, dry
CONCLUSIONS
sandstone and saturated sandstone. The P-wave velocity
measurements and UCS values are reproduced in Table 3 The P-wave velocity tests were conducted on 92 specimens,
and table 4 respectively. in which 47 of them can be classified as siltstone (or
siltstone with occasional beds of sandstone), while 45 of
Note: Three specimens failing at a very low load level due them can be classified as sandstone. Specimens of either
to the presence of pre-existing fractures were not included natural water content (dry condition) or in wet condition
in the present analysis. Because of the lack of sandstone were tested.
106
Table 4. Test results of uniaxial compressive strength (UCS) for siltstone and sandstone (Ma et al., 2010).
Siltstone Sandstone
Young’s Modules Young’s Modules
No. Stress (MPa) Strain (%) No. Stress (MPa) Strain (%)
(GPa) (GPa)
I* 171.09 0.28 65.41 AI* 293.17 0.38 79.97
I2 164.83 0.30 57.51 A2 341.90 0.44 80.73
I3 204.67 0.34 64.75 A3 298.90 0.41 73.56
I4 180.23 0.31 60.64 A4 223.34 0.38 71.60
I5 219.25 0.38 58.41 A5 192.31 0.26 75.04
Ave 188.01 0.32 61.34 Ave 269.92 0.37 76.18
SI1** 123.02 0.23 59.54 SA1** 232.02 0.23 76.18
SI2 138.45 0.28 61.13 SA2 241.19 0.37 75.31
SI3 128.67 0.28 61.81 SA3 298.43 0.43 76.79
SI4 127.01 0.24 56.04 SA4 231.67 0.35 75.77
SI5 125.89 0.26 62.29 SA5 271.47 0.41 74.83
Ave 128.61 0.26 50.16 Ave 254.96 0.38 75.74
*Dry condition *Dry condition
**Saturated condition **Saturated condition
The P-wave velocity measurements indicate that the P-

wave velocity in sandstone is generally higher than that
in siltstone. Based on the previous correlation by Kýlýç
and Teymen (2008) and other similar studies, the uniaxial
compressive strength is expected to increase with the P-
wave velocity measurement, i.e. sandstone is stronger than
siltstone. This finding agrees with our previous experimental
findings from UCS tests (Ma et al., 2010).
Secondly, the difference of P-wave velocity measurements in

the two measured orientations of siltstone, i.e. along bedding
planes and normal to bedding planes, is much larger than
that in sandstone. It indicates that the anisotropic nature
associated with the bedding in siltstone is more significant
than in sandstone.
Figure 4. UCS and P-wave velocity data of siltstone.
In order to establish a more significant statistical relationship
between the P-wave velocity measurement and the UCS
of rock cores, additional tests involving a bigger sample
size should be performed. This will become the scope of
the upcoming experimental study.
REFERENCES
[1] Kýlýç, A. and Teymen A., 2008. “Determination of mechanical

properties of rocks using simple methods”. Bulletin of

Engineering Geology and the Environment, 67(2), 237-244.
[2] Kahraman, S. and Yeken T., 2008. “Determination of physical
properties of carbonate rocks from P-wave velocity”. Bulletin
of Engineering Geology and the Environment, 67(2), 277-
281.
Figure 5. UCS and P-wave velocity data of sandstone. [3] Ma, G.W. and Wu, W., 2010. “Water saturation effects on
sedimentary rocks”. Civil Engineering Research (NTU), 23,
129-131.
[4] Ulusay, R. and Hudson, J.A., 2007. “The complete ISRM
suggested methods for rock characterization, testing and
monitoring”. 1974-2006.
107
ROLE OF FILLER IN MACRO STRUCTURE

OF ASPHALT MIXTURE AND ITS BINDING
CHARACTERISTIC WITH ASPHALT
Anggraini ZULKATI (angg0007@e.ntu.edu.sg)
WONG Yiik Diew (cydwong@ntu.edu.sg)
Darren SUN Delai (ddsun@ntu.edu.sg)
ABSTRACT: The role of filler in macro structure of asphalt mixture and micro interaction between filler and asphalt was investigated.
Preliminary results revealed that changing the amount of filler would affect aggregate packing structure, asphalt content and its properties,
and also workability during mixing and compaction, thus affecting asphalt mixture performance. Interaction of asphalt-filler mastic was
also examined using three types of fillers (granite, hydrated lime, kaolin) which revealed that the presence of filler in asphalt increased
the softening point and viscosity of asphalt. High-definition images of the morphology of the mastic were scrutinized that showed
evidences of “compatibility’ between asphalt and filler that could be attributed different affinity between filler and asphalt, or simply
micro voids around the filler from incomplete mixing process. It is conjectured that the presence of filler may cause stretching of atomic
bonds in the asphalt matrix, which either can enhance asphalt strength or possibly break the atomic bonds of asphalt matrix, leading to
undesired properties of mastic.
INTRODUCTION gradation design. Furthermore, interactions between asphalt

and filler which, jointly, may generate certain properties of
Asphalt mixture used in the surface layer of flexible the mastic. Different types of filler have different properties,
pavement road is formed from asphalt, coarse and fine may interact differently with asphalt, and hence create ‘new’
aggregates and filler. Aggregates are expected to provide properties in the asphalt-filler mastic. The interaction should
a strong stone skeleton to resist the repeated traffic load be dependent on physical and chemical properties of filler
applications. Asphalt provides adhesive action among as well as asphalt properties. As loading and temperature
aggregate particles. The mixture derives its strength from regime for asphalt pavement may vary, the micro interaction
the interlocking and frictional resistance of aggregates, between asphalt and filler should behave and perform
cohesion strength of the asphalt and adhesion binding differently at that variably condition. Thus, there is a need
between asphalt and aggregates to further examine the interaction between asphalt and filler
in micro level under different conditions.
Filler is the fraction of the mineral aggregate which mostly
passes the 75 μm sieve. Fillers fill voids between larger The on-going research was undertaken to investigate the
aggregates in the mixture. A good packing of coarse, fine role of filler in the macro structure of asphalt mixture. It
and filler shall provide a strong backbone for the mixture also investigated micro interaction among the constituent
(Vavrik et al. 2002, Alshamsi 2006, Qiu 2006, Rivera 2008). materials in asphalt mixture. Some preliminary findings
Filler may also alter properties of the asphalt, because the are reported in this paper.
filler can act as an integral part of the asphalt mastic which
is the combination of asphalt and filler.
MATERIALS AND METHOD
Even though fillers act as a ‘filler of voids’ or ‘asphalt
extender’, it is well documented that they play an important Experimental set up
role in providing strength stiffness and durability of asphalt

mixture (Kandhal et al.1998, Menglan and Chaofan 2008). The first task was focused on examining the filler effect with
The presence of filler in the asphalt mixture is even more respect to overall packing structure in aggregate gradation,
important because of their large surface area. Having its interdependence with other sizes of fine aggregate and
larger surface area, filler may absorb more asphalt and its its effect on workability during mixing and compaction.
interaction with asphalt may lead to different performances Six different gradations which vary in filler content were
of asphalt mixture (Kavussi and Hicks 1997, Taylor 2007, designed. Marshall samples were fabricated and tested for
Liao 2007, Lesueur 2009). mixture performance.
It is noted that despite the vast number of studies conducted The second task was to study the interaction of asphalt and
to design a good aggregate gradation, none attempted to filler in mastic and the resultant properties. Three different
types of fillers were selected to assess the effect of filler
comprehensively identify and investigate the role of filler in
in asphalt mastic.
108
Materials The Marshall stability tests were conducted according to

ASTM D6927-06 (ASTM, 2006). As for evaluating the
The type of aggregate used is Indonesian granite. Three material stiffness properties, the 5-pulse indirect tensile
fillers: granite, hydrated lime and kaolin were supplied by test as described in the Australian Standard Method 13.1
local chemical company. Sufficient quantity of aggregates (AS, 1995) was performed with the following conditions:
and filler was produced using a crusher and milling temperature at 25OC, peak loading force at 2kN, loading
machine. Asphalt Pen 60/70 supplied by Shell Bitumen period at 0.1 s, rest period at 0.9 s.
(Singapore) was used for asphalt mastic and asphalt
concrete mixture.
Asphalt-filler mastic preparation and test
Gradation design Asphalt-filler mastic samples were prepared to investigate

the property changes of asphalt in presence of filler. Filler
Mid value of Singapore’s standard W3B gradation and five was put into a 1000C oven for 24 hours to ensure moisture-
other gradations were designed as shown in Table 1. free particle surfaces. The asphalt-filler mastics were
produced by adding the correct mass of filler to the heated
Table 1. Gradation design. asphalt at a temperature of 150oC until a homogeneous
mastic was obtained. The mixing time was restricted to a
Coarse to fine % filler by
Gradation Gradation type maximum of 15 minutes. The mastic was then transferred
proportion (%) total mass
to make samples for further testing. Asphalt-filler mastic
W3B C-G 55:45 6
properties were then characterised based on the established
W3B-1 C-G 55:45 0 standard.
W3C-1 C-G 55:45 8
W3C-2 C-G 69:31 6 It is known that the specific gravity of three fillers is
different. The equivalent volume ratio was adopted for
W3D-1 F-G 40:60 4
calculating the concentration of different filler in asphalt.
W3D-2 F-G 40:60 6 The filler-asphalt ratios by volume equivalent are 0.06, 0.12,
C-G: coarse graded 0.18, and 0.24. The filler-asphalt ratio was determined by
F-G: fine graded dividing respective filler volume with asphalt volume.
W3B-1 is W3B gradation with exclusion of filler passing

75 μm sieve. W3C and W3D gradations are coarse and RESULTS AND DISCUSSION
fine-graded gradations respectively. Bailey method for
aggregate packing (Vavrik et al. 2002) was adopted Role of filler in macro structure
for designing these gradations. The method considers
the individual characteristics of aggregate properties. It The volumetric properties of each designed gradation
allows the adjustment of the aggregate voids by changing are presented in Table 2. Table 3 presents the Marshall
the packing of the coarse and fine aggregates and filler. stability and resilient modulus test results. The use of the
Aggregate packing ratios for each gradation were also Marshall stability and resilient modulus provides a basis
calculated to examine the packing structure of aggregates. for comparison of changes in material bearing capacity and
The aggregate ratios for all gradations were evaluated based stiffness respectively for different mixture types.
on Bailey packing ratios (Vavrik et al. 2002).
Table 2. Volumetric properties of samples at 5% binder content.
Density VMA VTM VFA
Asphalt mixture preparation Gradation
(g/cm3) (%) (%) (%)
W3B 2.34 13.92 3.47 75.08
Asphalt concrete mixture specimens were manufactured
W3B-1 2.29 15.20 5.20 66.85

using Marshall hammer compaction at 75 blows. The
Marshall specimens were fabricated with asphalt content W3C-1 2.32 14.95 4.31 70.61
of 5.0% based on the optimum asphalt content for W3B W3C-2 2.34 13.72 3.22 76.40
gradation. The 5.0% asphalt content may not be the W3D-1 2.24 15.95 6.34 60.27
optimum asphalt content for other gradations due to inherent
W3D-2 2.32 14.22 4.40 69.05
differences in terms of aggregate proportion. However, it
was chosen in order to analyse the volumetric properties VMA: voids in mineral aggregate
and mechanical performance of the mixture under the same VTM: voids in the mix
amount of asphalt. VFA: voids filled by the asphalt
After a 24-hour curing period for all Marshall specimens,

volumetric measurement for all specimens was conducted.
109
Table 3. Stability and resilient modulus at 5% binder content. As part of continuous size of gradation, filler content shall
Resilient relate to the proportion of other aggregates in gradation.
% filler by Marshall stability Once gradation of design is selected, the proportion of coarse
Gradation modulus
total mas (kN) and fine is then defined, and changing of the filler content
(MPa)
W3B 6 14.00 3,000 shall affect the proportion of aggregates in the fine fraction
only. Fines are small in size, but they have large surface
W3B-1 0 12.00 2,160
area, thereby playing an important role in determining the
W3C-1 8 12.70 2,850 asphalt content and the resultant mixture performances. As
W3C-2 6 16.50 2,990 seen in Tables 2 and 3, in comparing W3C-1 and W3C-2
W3D-1 4 11.70 1,800 properties or W3D-1 and W3D-2 properties, it was found
that changing a little amount of filler and the other fine
W3D-2 6 19.00 3,800
aggregates proportion impinged on all volumetric properties
and mechanical performance as well.
W3B-1 was simply designed by removing the filler
passing 0.075 mm in W3B gradation and replacing the
corresponding amount with particles at size 0.3-0.075 mm.
Role of filler in asphalt mastic
It can be seen that, although fillers are very tiny particle,
their presence affected the volumetric properties of the
The properties of three fillers are shown in Table 4. The
mixture. At certain amount, filler shall contribute to a
finest size is that for kaolin fillers, while granite filler is
better packing for the mixture by filling the voids created
the coarsest among the other two fillers. Hydrated lime has
by larger fine aggregates. The removal of that filler and
larger surface area, while granite has the lowest among the
replacement with other larger fine particles would cause
other two fillers.
‘overfilling’ of certain sizes, resulting in more voids (higher
VMA and VTM) and lesser density being created. With Table 4. Filler properties.
higher voids, VFA of that unbalanced mix would be lesser
than that of other mixes with better proportion of coarse, Apparent specific Specific surface
Fillers
fine and filler. gravity (g/cm3) area (m2/g)
Granite 2.56 0.758
With such filler-deficient volumetric structure, the stability hydrated lime 2.20 1.210
and stiffness of the mix were lower as compared with kaolin 2.48 1.110
other mixes with filler. Although the voids can be filled
by adding more asphalt binder, however, the aggregate Particle shape and morphology of the filler were examined
structure would not be well-packing. In addition, with poor using scanning electron microscopy (SEM). It was found
packing, having more asphalt can contribute to a lesser that the quartz in granite filler had largest and angular
shear and deformation resistance of the mix, particularly particles, hydrated lime had angular particles but they were
at high temperature. rather smaller than quartz, and kaolin particles were more
flaky with high aspect ratio.
The presence of filler in asphalt matrix may also affect
the workability during mixing and compaction. In simple Softening point and viscosity of asphalt-filler mastic at
mechanistic view, filler can perform as ‘a tiny roller different filler-asphalt ratios are shown in Figures 2 and
effect’ during mixing and compaction process (Figure 1). 3. It can be seen that at equivalent volume ratio, kaolin
This ‘tiny roller effect’, at certain filler asphalt ratio, shall filler provided highest effect in increasing the softening
correspond to a lesser friction, resulting in slightly faster point and viscosity of the mastic. It was found that the
and smoother re-orientation movement of larger aggregates stiffening effect of the filler in asphalt is dependent on
during mixing, thereby facilitating a tighter packing when filler size and surface area of filler. For industrial practice,
compaction load is applied. these mastic properties should be more appropriate for
batching design purpose. These properties may also more
appropriate in terms of establishing a correlation with the

resultant performance of asphalt concrete mixture.
As seen in Figures 4, 5 and 6, scanning electron microscopic

(SEM) images were used to examine the surface morphology
of the asphalt-filler mastic. The micro morphology was
used to approximately determine the compatibility of the
asphalt-filler mastic.
It was found that for the granite mastic, ‘white shadow’

which covers the granite filler in asphalt matrix always
Figure 1. Roller effect of filler. appear in the images. This shadow may indicate the
incompatibility between asphalt and filler which possibly
110
Figure 2. Softening point of asphalt-filler mastic.
Figure 5. SEM micrograph of hydrated lime mastic.
Figure 3. Viscosity at 135OC of asphalt-filler mastic.
occurs due to the different affinity of filler and asphalt or

simply indicative of small voids around the filler. Physically,
the presence of filler, especially at larger size, may cause
the stretching of atomic bonds of asphalt matrix, which Figure 6. SEM micrograph of kaolin mastic.
either can enhance asphalt strength or possibly break the
atomic bonds of asphalt matrix
CONCLUSIONS AND FUTURE RESEARCH
For lime and kaolin mastic, the white shadow was not WORKS
found. It can be due to the finesses of these two materials.
However, it was observed that some of lime and kaolin The experimental results demonstrated that filler contributed
fillers were not covered by asphalt. Incomplete mixing to the aggregate packing structure, and together with asphalt,
process might be one possible reason. The other reason jointly affected the workability, enhanced asphalt volume
can be due to the very fine size of these fillers, hence bulk and provided stiffening effect to the mastic.
asphalt may not ‘touch’ and desorbs into the filler.
It is emphasised that filler content should be designed to
consider the corresponding changes to other aggregate
proportion in the fine fraction. It leads to a reverse gradation
design, that is, once the optimum filler content is defined,
the proportion of other larger aggregates in the fine fraction
can then be designed. A well-balanced proportion of coarse,

fine and filler allows better packing with sufficient voids and
optimum asphalt needed, and provides desired performance
of the mixture in terms of strength, stiffness and durability.
It is noted that the properties of the asphalt-filler mastic
are also determined by interfacial bonding between the
bulk asphalt and filler. Such compatibility may result in an
undesirable balance of properties under different thermal
or load states.
With the benefits of preliminary findings, future tasks

Figure 4. SEM micrograph of granite filler mastic.
shall focus on: (i) investigating the role of filler in the
111
macro structure of asphalt mixture; (ii) examining asphalt- [6] Lesueur, D., 2009. “The colloidal structure of bitumen:
mastic characteristic at surface interaction level; and (iii) consequences on the rheology and on the mechanisms of
reengineering selected fillers, all with the aim of enhancing bitumen modification”. Advances in colloid and interface
the mastic and asphalt mixture performance. science, 145: 42-82.
[7] Liao, M., 2007. “Small and large strain rheological and
fatigue characterisation of bitumen-filler mastics”. PhD thesis,
University of Nottingham: UK.
REFERENCES
[8] Menglan, Z. and Chaofan, W., 2008. “Effects of type and
[1] Alshamsi, K.S., 2006. “Development of a mix design content of mineral filler on viscosity of asphalt mastic and
methodology for asphalt mixtures with analytically mixing and compaction temperatures of asphalt mixture”.
formulated aggregate structures”. PhD thesis, Louisiana State Transportation research record, 2051: 31-40.
University. [9] Qiu, Y., 2006. “Design and performance of stone mastic
[2] AS 1995. Method 13.1: Determination of the resilient asphalt in Singapore conditions”. PhD thesis, Nanyang
modulus of asphalt – Indirect tensile test, Australian standard Technological University: Singapore.
AS 2891.12.1. Methods of sampling and testing asphalt. [10] Rivera, F.A., 2008. “Evaluation of the Bailey method as a
Australia. tool for improving the rutting resistance of mix designs using
[3] ASTM 2006. D1559-89: Standard test method for resistance New Hampshire aggregate”. MSc thesis, University of New
to plastic flow of bituminous mixtures using Marshall Hampshire.
apparatus. Annual books of ASTM standard, 04.03: 1103- [11] Taylor, R., 2007. “Surface interactions between bitumen and
1187. Philadelphia. mineral fillers and their effects on the rheology of bitumen-
[4] Kandhal, P.S., Lynn, C.Y. and Parker, F., 1998. “Characterization filler mastics”. PhD thesis, University of Nottingham: UK.
tests for mineral fillers related to performance of asphalt [12] Vavrik, W.R., Pine, W.J., Carpenter, S.H. and Bailey, R., 2002.
paving mixtures”. NCAT report 98(2). “Bailey method for gradation selection in hot-mix asphalt
[5] Kavussi A. and Hicks, R.G., 1997. “Properties of bituminous mixture design”. Transportation research board, Washington
mixtures containing different filler”. Proceedings of the D.C.
association of asphalt paving technologists, 66: 153-186.
112
STRUCTURES AND MECHANICS
A CO-ROTATIONAL SHELL ELEMENT

WITH MATERIAL NONLINEARITIES
Xu Jin (xu0003in@e.ntu.edu.sg)
Tan Kang Hai (ckhtan@ntu.edu.sg)
Lee Chi King (ccklee@ntu.edu.sg)
ABSTRACT: This paper shows a 6-node curved nonlinear co-rotational shell element with nonlinear models. Vectorial rotational variables
are used in this co-rotational element, so that a symmetric stiffness matrix can be obtained and the updating procedure is simplified. In
order to investigate the behaviour of structures with nonlinear material constitutive relationships, a layered model is adopted so that material
properties varying along the thickness of shell structures can be simulated independently. Since the co-rotational approach is adopted,
the material constitutive model based on stress-strain relationship can be used directly in each elemental local coordinate system.
INTRODUCTION node i: ngi = [pi,X, pi,Y, pi,Z], whose modulus are the two
smaller ones.
Co-rotational (CR) approach is an efficient method to solve
geometrical nonlinearity, which finds its origin in the paper
of Wempner (Wempner 1969). The popularity of the CR
formulation is largely due to the decomposition of a large
displacement into rigid body motion and deformational
displacement. In this way, a geometrically-nonlinear problem
is transferred to be a small strain problem and different
material nonlinearities are easier to be implemented in the
CR formulation, compared with Total Lagrangian (TL)
formulation or Updated Lagrangian (UL) formulation. Figure 1. The global and the local coordinate system.
CO-ROTATIONAL FORMULATION The relationship between the nodal variables in the global
and the local systems is given by Equation (1) and (2).
Coordinate Systems
ti = R(Xi – X1) – R0(Xi0 – X10) (1)
Three coordinate systems are defined in the present CR
elements: 1) the global system, 2) local systems and ri0 = Rh0ngi0, ri = Ringi (2)
3) natural systems. The global system OXYZ is defined
during the modelling stage by users. Each element has where R and Rh are the rotation matrices for translational
its own local system. The local system is fixed on the and rotational displacements, respectively; ngi (i = 1, 2, …,
element and co-rotates with the element. The rotation and 6) is the normal vector for node i; the subscript 0 refers
translation of a local system represents the rigid body to coordinates or normal vectors that are in the initial
rotation and rigid body translation of its corresponding configuration.
element, respectively.
Kinematics of the Shell Element

Nodal Degree of Freedom

The in-plane strain ε and the shear strain γ for shallow
In the present shell element, there are five degrees of curved shell element are given in Equation (3) and (4),
freedom for each node, namely, three translational variables respectively.
and two rotational variables. The three translational variables
include diT = [Ui, Vi, Wi] for node i in the global system ε = εm + zχ (3)
and tiT = [ui, vi, wi] for node i in the local system, while
the two rotational ones are vectorial rotational variables
(Li & Izzuddin et al. 2008; Xu & Tan et al. 2010), which (4)
are piT = [pi,mi, pi,ni] for node i in the global system and
riT = [ri,x, ri,y] for node i in the local coordinate system. where εm is the in-plane strain of the reference surface
pi,mi and pi,ni are the components of the normal vector at (the mid-surface of the shell structure), which is given by
113
Equation (5); χ is the curvature of the element, which is Stiffness Matrix and Internal Force Vector
given by Equation (6); z is the z-coordinate of the point
of interest in the local system; the nabla symbol ∇ refers The strain energy of an element is given by Equation
to the gradient of a scalar function. (8).
(8)
where Vj is the volume of the jth layer; V is the volume of the

whole element; We is the work done by the external force;
D1 (size: 3×3) is the material constitutive matrix for in-plane
(5) stress-strain and D2 (size: 2×2) the material constitutive
matrix for shear stress-strain. For linear elastic material, D1
and D2 is given by Equation (9); for elasto-plastic material,
(6) D1 is modified by elasto-plastic modulus Dep.
The derivative ∇f (x, y, z) = J-1∇f (ξ, η, ζ), where J is the

Jacobian matrix [∂x/∂ξ], is applied in Equation (4) ~ (6).
Layered Model ,
(9)
In order to incorporate material nonlinearity, a layered model
is adopted in the present shell element so that material where E is the Young’s modulus; μ is the Poisson’s ratio;
constitutive relationships can be modelled in each layer. k is the shear correction parameter, k = 5 / 6.
The layers are numbered sequentially, starting from the
bottom surface of the element and each layer may have The stiffness matrix kTL (size: 30×30) and the internal
different thicknesses, as shown in Figure 2. force vector fLT (size: 30×1) of the element are the Hessian
matrix and Jacobian matrix of strain energy.
The internal force vector in the global system fG is given

by Equation and the stiffness matrix in the global system
kTG is given by Equation (11).
fG = TTfL (10)
(11)
Figure 2. Layered model. where T is the transformation matrix: T = [∂uL / ∂uGT].
The in-plane strain εj at the mid-surface of the jth layer is

given by Equation
MATERIAL NONLINEARITY
Elasto-plastic Constitution
εj = εm + zjχ (7)
Since the layered model is adopted, the modification for
where zj is the distance from the mid-surface of the element
material nonlinear analyses is focused on the material
to the mid-surface of the jth layer. It should be noted that
constitutive matrix. In the present element, only in-plane

the in-plane strain εm may not be the pure membrane strain
stress components are taken into account when dealing
for material nonlinear cases.
with elasto-plastic material.
The shear strain is assumed to be uniformly distributed
The incremental stress is given by Equation (12)
along the thickness direction of the element, which means
that the shear strain of each layer is the same and is given
(12)
by Equation (4).
where H is the hardening parameter and vector a is the
first derivative of effective stress with respective to the
stress components.
114
Reinforced Concrete Model
The reinforced concrete model employed in the present shell

element is proposed by Owen and Figueiras (1984).
NUMERICAL EXAMPLES
Roll-up of a Clamped Beam Figure 4. The computational model of the flat plate.
An initially flat shell is fully fixed at one end and is acted by

a bending moment at the other end. The analytical solution
for this example is governed by the classical formula:
(13)
where M is the bending moment applied on the beam

and ρ is the curvature of the beam. For M = 2π EI / L,
where L is the length of the beam, the beam rolls up into
a complete circle.
Figure 5. The equilibrium path of the flat plate.
area to the loading point, the more refined is the mesh. For
both meshes, 5 layers for each element are used.
A Reinforced Concrete Slab
A corner supported square plate is loaded by a concentrated

force at the centre, as shown in Figure 6.
Figure 3. The process of rolling-up for the beam.
Table 1. The rotation angle for the cross-section at the free end.
λ 0 0.1 0.2 0.3 0.4 0.5
θ(rad) 0.000 0.692 1.226 1.803 2.415 2.982
λ 0.6 0.7 0.8 0.9 1.0
θ(rad) 3.572 4.213 4.864 5.620 6.336
The geometric and material properties for the beam are

chosen as L = 12, I = bd3 / 12 = 1 / 12, E = 2.40 × 105,
μ = 0.3. A total of 20 elements (1 × 10 × 2) is used for
Figure 6. The computational model of the RC slab.
this example.
Three cases (S1, S2, S3) are tested with different amounts
A Steel Plate Applied by a Point Load of reinforcement in each direction, but with the total amount
of reinforcement in the plates kept constant. The ratio of
A perfectly elasto-plastic analysis of a simply supported the amounts of reinforcement in the two direction (ρX : ρY)
square flat plate is conducted in this example (Figure 4). is shown in Table 2.
The geometric and material properties for the computational Table 2. The reinforcement for the three cases in the RC slab.
model are chosen as: the length L = 16 m, thickness t = slabs S1 S2 S3
0.5 m, Young’s modulus: E = 7.05 × 106 N/mm2, yielding
ρx : ρ Y 1:1 1:1.89 1:2.75
stress: σ0 = 2.44 × 102 N/mm2. The computational model is
meshed by 200 elements (10 × 10 × 2) and 800 elements
The geometrical properties of the computational model and
(20 × 20 × 2). For the case of 200 elements, the closer the
the mesh used are shown in Figure 6. The configuration
115
CONCLUSIONS
A 6-node curved co-rotational shell element with layered

model is presented. The present element has several
features: 1) the rotational variables are commutative as
Figure 7. The concrete layers and reinforcement layers
the translational variables, so that the stiffness matrix of
along the thickness direction.
this element is symmetric; 2) the updating of the rotational
variables are simplified, because the spin matrix for the
of the cross-section is shown in Figure 7. The material updating of the convention rotational variables is not
properties for concrete and reinforcement are listed in needed; 3) layered model is employed in this element
Table 3. so that it can be used in material nonlinear analyses, for
example: reinforced concrete structures, elasto-plastic
Table 3. Material properties for RC slab (kN, cm). materials and so on.
Concrete
Young’s modulus Ec = 1640
Poisson’s ratio μ = 0.0
REFERENCES
Ult. Comp. Strs. fc’ = 4.30
[1] Wempner, G. 1969. Finite elements, finite rotations and small
Ult. Ten. Strs. ft = 0.3 strains of flexible shells. International Journal of Solids and
Ult. Comp. Strn. εu = 0.0035 Structures 5(2): 117-153.
Ten. Stiff. Coeff. α = 0.7 [2] Li, Z.X., Izzuddin, B.A., et al. 2008. A 9-node co-rotational
Ten. Stiff. Coeff. εm = 0.0035 quadrilateral shell element Computational Mechanics 42(6):
873-884.
Reinforcement
[3] Xu, J., Tan, K.H., et al. 2010. A Co-Rotation Shell Element
Young’s modulus Es = 20100 with Material Nonlinearities. In B. H. V. Topping, J. M.
Young’s modulus ET = 700 Adam, F. J. Pallarés, R. Bru and M. L. Romero (ed.), The
Poisson’s ratio μ = 0.0 Tenth International Conference on Computational Structures
Yield stress σu = 60.0 Technology, Valencia, Spain, Stirlingshire, Scotland: Civil-
Comp Press.
Ult. Stress. σu = 70.0
[4] Owen, D.R.J. and Figueiras, J.A. (ed.) 1984. Finite element
software for plates and shells. Swansea, United Kingdom,
The results are plotted in Figure 8 together with the results Pineridge Press.
from numerical analysis (Owen & Figueiras 1984) and
[5] Duddeck, H., Griebenow, G., et al. 1978. Material and time
experiment (Duddeck & Griebenow et al. 1978). dependent nonlinear behaviour of cracked reinforced concrete
slabs. G. Mehlhorn, H. Ruhle and W. Zerna. Darmstadt, West
Germany.
Figure 8. The equilibrium path of the RC slab.

116
COLLISION ANALYSIS OF OFFSHORE

FLEXIBLE RISERS
Jian Wen He (heji0006@e.ntu.edu.sg)
Ying Min Low (ymlow@e.ntu.edu.sg)
ABSTRACT: Flexible marine risers are compliant to external forces from waves, current and platform motions, and clashing between
risers is an important concern. In deepwater developments where the number of connected risers is large, it is not economical to space
them too far apart. In this regard, it is necessary to establish the probability of riser clashing throughout the service life; however, at
present there appears to be no systematic procedure for assessing this risk. This article presents a novel procedure for estimating the
probability of riser clashing based on post-processing results obtained from time domain simulations of flexible risers subjected to
random wave loads.
INTRODUCTION Riser clashing is usually controlled within a low level, for

example, less than 10-4 during field life. It is not appropriate
As the emphasis of the oil and gas industry shifts towards to use experimental methodology for assessment, and
greater water depths, deepwater marine risers play an simulation is a better alternative. However, in literature,
indispensable role in the production of hydrocarbons. there is no approach to offer a quantitative conclusion to
Several major types of deepwater risers are Flexibles, show the reliability of a riser system.
Hybrid Riser Towers (HRTs) and Steel Catenary Risers
(SCRs); Top Tensioned Risers (TTRs) and umbilicals are In view of current gaps, the motivation of this study is to
also widely in use. Among the types mentioned above, establish a probabilistic approach to assess the reliability
flexible risers seem to be popular around the world, and of riser systems in which riser clashing rarely occurs.
they dominate the oil and gas production in Brazil. In Uncoupled time domain analysis of the riser will be
Europe, the UK section of the North Sea alone has more performed as a first step. It is sensible to exclude VIV
than 1000 flexible risers. and wake effects at this preliminary stage. For simplicity,
current will not be included in the analysis.
In deepwater fields where the number of connected risers
tends to be large, designers generally prefer compact riser
systems for economical reasons. However, the compliancy PREDICTION OF EXTREME RESPONSE
of flexible risers means that close riser spacing will increase
the risk of mutual contact and collision. The two conflicting General
design requirements bring about the need for an optimized
riser layout design, which should have certain reliability Marine risers are usually analyzed as line elements using
against riser clashing. In fact, riser clashing is a complicated a numerical model such as the popular finite element
problem, because both static contact and dynamic clashing method. In this article, the commercial software is used
can cause the risers to have fatigue or coating damage, and which performs the time domain analysis using the lumped
the latter may even lead to loss of integrity. To address this mass approach. The time history of the nodal positions is
concern, DNV code (Det Norske Veritas, 2009) recommends extracted directly from the simulation results. It is worth
that the minimum clearance between two risers be the noting that the approach in this article can be applied to any
sum of their diameters without considering safe factors. other numerical model provided that all the nodal positions
Nevertheless, when optimization is of concern, engineers are available at all time steps. Here, time history X(t) is
prefer to know the reliability of a particular system rather recommended to be the maximum normalized clearance
than being given inflexible criteria. Riser dynamics is so between two risers. The generation of X(t) is discussed at
varied and complex that recommendations based on past length by He and Low (2010).
experience may not be suitable for a particular problem.
For one thing, nonlinearity of line structure evidently In this approach, it is assumed that the critical condition
complicates the prediction of its response. For another will occur during a typical storm according to a 100-year
thing, vortex-induced vibrations (VIV) and wake effects return period, and the sea state is stationary for typically
between two close risers are still too intricate to be fully three hours. A wave spectrum is chosen to express features
understood. Such effects are significant; for example, VIV of short-term conditions.
is able to cause large deflection in the risers to the extent
of one diameter. Crossing rate analysis is the method adopted in this article.
In time domain simulation for a stationary sea state, the
117
crossing rates are commonly extracted from the time history

X(t). If the total simulated duration is T0, which usually
includes a number of independent simulations with duration
T, and the associated number of up-crossings exceeding
threshold z within [0, T0] is , then the mean up-
crossing rate is
nz+ (0; T0)
vz+ (0; T0) = m …(1)
The probability of crossing a given threshold within a

certain time range can be estimated as
Pf = 1 – exp (–vz T) …(2)
Here, Pf represents the probability for X(t) to exceed

threshold z in duration T and Eq. (2) is a good approximation
if the crossing level z is sufficiently high.
Figure 1. Riser model.
Naess method
symmetrical plane of two risers (i.e. in the y-direction) and
assigning a small top end clearance (0.6 D) and a small
Naess et al. (2008) gave an empirical formula of the
separation angle (0.5 degrees). The significant waveheight
crossing rate
Hs and mean zero crossing period T0 are 15.7 m and 13.5
s separately for a severe JONSWAP sea state. All the
vz = q(z)exp{[–a(z – b)]c} …(3)
properties and structural details are reported in Tables
1-2.
where a, b and c are constant parameters, and q(z) can
also be regarded as constant when z is in a high level. q
A wave time history derived from a certain number of
is recommended to be estimated by the average of
wave components is synthesized. The wave spectrum is
for the largest values of z, where fz is the response pdf. b
divided into 250 components with equal frequency interval
is determined empirically, it should not be larger than the
approach from 0.07 Hz to 0.239 Hz. To improve the quality
smallest z for the fitting and the choice of b must optimize
of prediction, thirty independent random number sets are
the curve fitting (Naess et al., 2008). It is the first time
employed to simulate 30 one-hour storms. To make the
to apply this method to a riser system. Eq. (3) can be
computational intensity remain in an acceptable level, it
rearranged to give
is sensible to simplify the environmental loadings. Thus,
v only the first-order wave forces are considered in the model
1n (–1n( zz )) = 1n(a) + c(1n(z – b)) …(4)
and the current, VIV and wake effects are neglected at this
incipient stage.
In Eq. (4), the left side is treated as the dependent variable,
1n(z – b) and c are regarded as independent variable and its
Table 1. Structure of risers.
gradient. Because 1n(a) is a constant, q and b are estimated
beforehand. The art of linear fitting can be directly applied Section
Line Type Descriptions
to achieve parameters a and c. The desired crossing rate Length (m)
obtained from Eq. (3) can be put back to Eq. (2) to attain Flexible (upper section) 605 Included in analysis
the corresponding Pf. 80 Not included in
Flexible with buoyancy
analysis
341 Not included in

Flexible (lower section)
MODEL AND NUMERICAL RESULTS analysis
Total 1026
Model
A simple model, which comprises two flexible risers

following the Lazy Wave configuration, is set up in the
commercial software Orcaflex in Fig. 1. The two ends of
a riser are attached to the floater and seafloor with pinned
joints respectively. To alleviate the heavy burdens of
calculation, only around 60% of the length (95 elements) is
contained in the analysis. A severe condition is elaborated
by allowing the wave direction to be perpendicular to the
118
Table 2. Properties of risers.

Riser Types
Flexible with
Flexible
Buoyancy
Outside Diameter (mm) 478.4 873.0
Wall Thickness (mm) 133.1 330.4
Mass (kg/m) 200.0 431.1
Area (cm2) 1443.9 5632.1
EI (kN×m2) 153 153
Cd (~) 1.200 0.979
Hang-off Angle (°) 10 N.A.
Content Density (kg/m3) 500 500
Numerical results
The up-crossing rates are directly extracted from a 30-hour Figure 3. Crossing rate extrapolation by Naess method;
simulation time history. Insufficient exceedances of certain 1n(–1n((vz+)/q)) is plotted against 1n(z – b); q=0.36, b=-0.05,
critical values will produce unrealistic crossing rate. In a=14.14, c=1.097; range of value from z = 0.15 to 0.35;
Fig. 2, the unsteady pattern at the tail is mainly due to
lack of data. CONCLUSIONS
Currently, the focus on ultra-deep water field operation

and compact riser designs makes riser collision a topic of
concern. As a safety precaution, a riser system is always
designed in a conservative way to avoid riser clashing.
This article has proposed an effective approach to assess
the low probability riser clashing. Such an approach is
based on the existing nodal positions obtained from a
time domain simulation and does not require extra work.
For the riser system in which no collision is allowed, the
proposed approach can directly calculate the probability of
failure in terms of riser clashing.
REFERENCES
[1] Det Norske Veritas, 2009, Offshore Standard DNV-RP-F203:

Riser Interference.
[2] He J.W. and Low, Y.M., 2010. “Probabilistic Assessment of
Figure 2. Crossing rate. the Clashing between Flexible Marine Risers”. Proc. OMAE,
Paper No. 20046, Shanghai.
[3] Naess, A., Gaidai, O. and Teigen, P.S., 2008. “Extreme response
Naess method is robust and gives a practical guide for the prediction for nonlinear floating offshore structures by Monte
extrapolation. It is worth noting that, to achieve an accurate Carlo simulation”. Appl. Ocean Res., 29, pp. 221-230.
extreme value distribution, the up-crossing events of the

high response levels must be statistically independent.
This requirement is usually satisfied provided that the total
damping of a system is small such as, typically, a riser
system. It is shown in Fig. 3, Pf = 0.13% at the desired
threshold z=1. The crossing rate is extrapolated from
=2.4E-3 to the expected value about 1E-7 which is
expensive in CPU time with direct simulations.
119
EFFECTS OF ANISOTROPIC
PERMEABILITY OF FRACTURED ROCK
MASSES ON ROCK CAVERNS
Sun Jianping (sunjp@ntu.edu.sg)
Zhao Zhiye (czzhao@ntu.edu.sg)
ABSTRACT: An isotropic assumption is often applied to analyze in-situ permeability tests of fractured rock masses, and the isotropic
hydraulic conductivities are then used directly in the seepage analysis. However, the hydraulic conductivities are normally anisotropic
in fractured rock masses and the effects of the anisotropic permeability should be taken into account in rock engineering analysis,
especially for seepage analysis of underground oil storage caverns. In this study, an underground oil storage cavern project is analyzed
and the Oda’s method is used to determine the anisotropy in permeability. The anisotropy in permeability is determined using the fracture
orientation and the in-situ stress information from the field survey. A typical cavern unit is numerically modeled using the computer
code FLAC. The effects of anisotropic permeability on water pressure and critical gas pressure are studied. The results indicate that the
calculated results based on the in-situ hydraulic tests with isotropic permeability assumption can be used safely in the underground oil
storage cavern project.
INTRODUCTION (v) there is no head loss at intersections between fractures.

Based on these assumptions, Oda (1985) proposed hydraulic
The basic principle of oil storage in unlined underground conductivity components as follows:
caverns is that the hydraulic potential in the rock mass
around the caverns should be higher than the potential kij = λg(Pkkδij – Pij) / μ …(1)
on the perimeter of the storage caverns. The groundwater
flow through fractured rocks should be analyzed carefully Where
for underground storage cavern, so as to make sure no πρ ∞ ∞
leakage under various operating conditions. For analyzing Pij = 4 ∫0 ∫0 ∫Ω r 2 t 3 n i n i E(n,r,t)dΩdrdt …(2)
the groundwater flow correctly, some borehole hydraulic
tests were performed around the project site. Based on the and
isotropic assumptions, the hydraulic conductivities were
obtained from the correlation curve of the injected tests. Pkk = P11 + P22 + P33 …(3)
However, hydraulic conductivities are usually anisotropic
in fractured rock masses, so the question of whether where g (LT-2) is the gravitational acceleration; μ(L2T-1)
the analysis results based on in-situ hydraulic tests with is the kinematic viscosity; λ is a nondimensional scalar
an isotropic assumption can be used in the preliminary dependent on the connectivity among joints, and can be
design of underground oil storage cavern projects should set to 1/12 for practical applications; δij is the Kronecker
be addressed. delta; ρ is the number of joints per unit volume; ni is the
component of n projected on the orthogonal reference
axis system (xi =1,2,3); n is the fracture orientation; r
ANALYSIS OF ANISOTROPIC PERMEABILITY is the fracture length; and t is the fracture aperture. If
BASED ON IN-SITU DATA the unit vectors n normal to the fractures are orientated
inside a small solid angle dΩ around n and the diameters

Theoretical model and the apertures range from r to r + dr and from t to t
+ dt, respectively, E(n,r,t)dΩdrdt is then given as the
If a fractured rock mass block can be treated as a probability of the unit normal of (n,r,t) fractures.
homogeneous, anisotropic and porous medium, it would
obey Darcy’s law. In order to obtain the components
of the hydraulic conductivity in fractured rock masses, Site investigation data
the following assumptions are made: (i) each fracture is
idealized by a set of parallel plates with a uniform aperture As part of the site investigation program, six vertical
t; (ii) the solid matrix is impermeable; (iii) the hydraulic boreholes are drilled to investigate rock hydraulic properties.
gradient is uniformly distributed over the whole body; (iv) The locations of six vertical boreholes, B1, B2, B3, B4, B5
seepage flow through a fracture can be treated as a laminar and B6, are shown in Fig. 1. The fracture orientation data
flow between parallel plates with a uniform aperture; and and dip/dip angle are obtained from the borehole survey.
120
S h, S H and S v (MPa)
0 2 4 6 8 10
80
100
120 SH
Depth (m ACD)
Sh
140 Sv
160
180
200
Figure 3. Stress’s profile for borehole B5.
In-situ anisotropic permeability analysis
In additional to the geostatic stress and fracture orientation

information, fracture length and aperture information should
be known in order to obtain the hydraulic conductivities.
Figure 1. Schematic map of locations of vertical boreholes
for the underground cavern. But the site investigation does not provide such information,
so some assumptions are needed. Let us assume that the
statistical variables n and r are mutually independent. A
In total, 72 hydraulic conductivity measurements were lognormal distribution function is used for the distributed
conducted in the six boreholes, by the injection tests. Fig. forms of fracture length.
2 presents the measured hydraulic conductivity data at the
six boreholes at the depth between -40 mACD and -200 Based on Oda’s theoretical model and in-situ investigation
mACD, which ACD is the abbreviation of “Admiralty Chart data, the ratio of anisotropy in permeability defined as
Datum”. The results show that the hydraulic conductivity k1 / k3 and k1 / k2 can be determined, in which k1 , k2 and k3
varies between 10-11 m/s and 10-4 m/s. It should be noted that are the principal hydraulic conductivities. Fig. 4 shows the
the hydraulic conductivities obtained from the correlation calculated results of the ratios of anisotropy in permeability
curve of injected water pressure and flow quantity of the for six vertical boreholes between -80 mACD and -180
injected water are based on the assumptions that the fracture mACD depth. It can be seen from Fig. 4 that the ratio of
rock is homogeneous, isotopic and porous media and the anisotropy in permeability decreases with increased depth,
flow geometry is cylindrical, which do not reflect the due to the effect of geostatic stress. In borehole B1, the ratio
anisotropic property of fractured rock. In order to derive of k1 / k3 decreases from 1.91 to 1.65 with the increased
the local stress regime at the proposed development area depth. In borehole B2, the ratio of k1 / k3 decreases from
for the oil storage cavern, a total of 10 hydraulic and 2.50 to 2.09 with the increased depth. The ratios at the
hydrofrac/hydraulic injection tests were conducted in the other boreholes are between the ratios at B1 and B2. The
uncased section of borehole B5 between -96 mACD and results also show that the ratios k1 / k2 for all the boreholes
-181.4 mACD (Fig. 3). are close to 1. The calculated results illustrate that k1 and
k2 are almost in horizontal directions and k3 is close to
Hydraulic conductivity (m/s)
10
-13
10
-11
10
-9
10
-7
10
-5
10
-3 vertical direction. Hydraulic conductivities obtained from
the correlation curve of the injected water pressure and flow
-40
quantity of the injected water are based on the assumptions
that the fracture rock is homogeneous, isotropic and porous

-80 media and the flow geometry is cylindrical. As k1 and k2
are almost the same and along the horizontal direction, the
Depth (m ACD)
-120 hydraulic conductivity measured from the injection test

(Fig. 2) can be considered as the average value of k1 and
-160 k2, and k3 can be estimated by the ratio of k1 / k3 (Fig. 4)
B1
from the Oda’s method.
B2
-200 B3
B4
B5
B6
-240
Figure 2. Hydraulic conductivity with increasing depth

at six boreholes.
121
B1
conductivity obtained from the injection test (Fig. 2) is used
-60
B2 as the horizontal hydraulic conductivity kH. The vertical
B3
-80 B4 hydraulic conductivity kV is estimated based on Oda’s model
B5
B6
in Section 2. The ratios of kH / kV equal to 1.0, 1.5, 2.0 and
2.5, respectively, will be used in the following analyses
Depth (m ACD)
-100
to study the effects of anisotropic permeability on the
-120 storage caverns.
-140
The groundwater flow is assumed to be steady in this study.
Constant pressure which is equal to 15m water column
-160
is used on the upper boundary to simulate the effect of
-180 sea water pressure. Constant head boundaries at the water
1.4 1.6 1.8 2 2.2 2.4 2.6
curtains are assigned with the fixed hydraulic potential H
K 1/K 3 equals to 0 mACD. Two different phases, construction phase
Figure 4. The ratios of anisotropic permeability and operation phase, are considered. The main differences
for six vertical boreholes. between the two cases are related to the pressure inside
the cavern and water gallery condition. In the construction
phase, the caverns and water galleries are filled by air with
NUMERICAL MODELING OF UNDERGROUND the atmosphere pressure of 0 MPa. In the operation phase,
OIL STORAGE CAVERNS the water galleries are filled by water and the fixed hydraulic
potential H which is equal to 0 mACD is applied as the
A typical oil storage cavern unit is modeled as a 2D boundary condition; the caverns are filled with oil in density
model. The computer code FLAC is adopted to model = 900 kg/m3 and the distribution of hydraulic pressure at
the groundwater flow into the caverns. The crowns of the wall of the cavern equals to the oil pressure.
caverns are located at -119 mACD. The shape of the
cavern is horseshoe, with the height = 26m and the width
= 20m. The vertical water curtain is installed between the INFLUENCES OF ANISOTROPIC
contiguous storage units and connects with water gallery.
PERMEABILITY ON UNDERGROUND
The modeling area is 140m wide and 165m deep. The mesh
size is 2.5m×2.5m. The distance between the two caverns STORAGE CAVERNS
is 40m. The representative cross section for the model is
shown in Fig. 5. Regarding rock mass characteristics, the Influences on water pressure distributions
four main hydrogeological classes have been taken into
account along the vertical direction: the recent sediment/ The calculated results show that the water pressure between
landfill class, the weathered zone class, the low confined the two caverns decreases as the ratio of kH / kV increases.
zone class and the confined bedrock class. During the construction phase, the maximum water pressure
changes from 148kPa to 75.1kPa as the ratio of kH / kV
JOB TITLE : (*10^2) increases from 1.0 to 2.5. During the operation phase,
FLAC (Version 5.00)
the maximum water pressure changes from 257.13kPa to
LEGEND -0.200
185.30kPa as the ratio of kH / kV increases from 1.0 to 2.5.
7-Aug-09 12:51
step 0 Hence, a conservative result will be obtained by using
-1.100E+02 <x< 1.100E+02
-2.075E+02 <y< 1.250E+01
-0.600 isotropic assumption to assess the rock stability between
Permeability
7.136E-12
4.077E-10
the caverns, as a higher water pressure will reduce joint
5.097E-09
1.019E-07 -1.000
strength, and reduce the rock mass strength.
Grid plot water water
gallery gallery
0 5E 1
water
curtain
water
curtain
The results also show that the water pressure distribution
between the vertical water curtain and the cavern has little
-1.400
variation as the ration of kH / kV increases. This means that

-1.800
the anisotropic permeability has little influence on the rock
stability between the vertical water curtain and the cavern.
The vertical water pressure distribution above the cavern
-0.800 -0.400 0.000 0.400 0.800
(*10^2)
Figure 5. A numerical model for the rock cavern unit. crown is studied. As depth changes from -15 mACD to
-119 mACD, the water pressure increases first and then
decreases. It can be seen that an increase of water pressure
The rock classes of landfill/recent sediment, the weathered above the cavern crown will happen as the ratio of kH / kV
rock and the low confined zone are assumed to be isotropic increases. During the construction phase, the water pressure
due to their adequate weathering. Anisotropic permeability at -108 mACD increases from 289.73kPa to 316.00kPa
is considered for the confined bedrock class only. Because as the ratio of kH / kV increases from 1.0 to 2.5. During
k1 and k2 are almost the same and in the horizontal direction the operation phase, the water pressure at -108 mACD
in the confined bedrock, the geometric average hydraulic increases from 344.30kPa to 423.38kPa as the ratio of
122
kH / kV increases from 1.0 to 2.5. This implies that the rock and the in-situ stress obtained from field survey. The results
stability around the cavern crown decreases as the ratio of show that the ratio of k1 / k3 varies from 1.65 to 2.50, and
kH / kV increases. k1 and k2 are almost in the horizontal direction and k3 is
close to vertical direction in this site.
Influences on critical gas pressure A typical cavern unit is numerically modeled using the
computer code FLAC. The geometric average hydraulic
The gas tightness of rock caverns should also be considered. conductivities obtained from the injection test are used as
Goodall et al. (1988) recommended a practical design the horizontal hydraulic conductivity kH and kH / kV equal to
criterion that no gas will leak as long as the water pressure 1.0, 1.5, 2.0 and 2.5, respectively, are used in this numerical
increases along all possible gas leakage paths away from model. The effects of anisotropic permeability on water
the caverns. According to Goodall’s criterion, at the critical pressure and critical gas pressure are studied carefully.
gas pressure, the groundwater pressure in the vicinity of The results show that most calculated results which are
the cavern is equal to the gas pressure at some point on based on in-situ hydraulic tests with isotropic permeability
the boundary of cavern, i.e., ∂p / ∂n = 0, where p is the assumption can be used safely in the underground oil
groundwater pressure and n is the unit normal vector at storage cavern project.
this point. Liang and Lindblom (1994) suggested the “critical
gas pressure” as the maximum tolerable gas pressure for Only the anisotropy induced by fracture orientation and
a given storage system at no gas leakage conditions. In in-situ stress are considered in this paper. The influences of
this study, the Goodall’s criterion is used to determine bolting, shotcreting and redistributed stress after excavating
the critical gas pressure and the influence of anisotropic on anisotropy in permeability and heterogeneous property
permeability on critical gas pressure is investigated. The of fractured rock mass are not studied. Further studies
calculated results show that an increase of the ratio of that consider these influences will be carried out in our
kH / kV can linearly increase the critical gas pressure. The future work.
critical gas pressure is 0.846 MPa and 0.890 MPa for
kH / kV = 1.0 and kH / kV = 2.5, respectively. Using the
isotropic assumption, a safe critical gas pressure is obtained REFERENCES
and can be used to design the engineering project.
[1] Oda, M., 1985. “Permeability tensor for discontinuous rock
masses”. Geotechnique, 35(4), 483-495.
CONCLUSIONS [2] Goodall, D.C., Aberg, B. and Brekke, T.L., 1988. “Fundamental
of gas containment in unlined rock caverns”. Rock Mechanics
In this study, Oda’s method for determining the anisotropic and Rock Engineering, 21, 235-258.
permeability, which is difficult to achieve from the in-situ [3] Liang, J. and Lindblom, U., 1994. “Analyses of gas storage
tests, is used for the seepage analysis of an underground oil capacity in unlined rock caverns”. Rock Mechanics and Rock
storage cavern. In this method, anisotropic permeability of a Engineering, 27(3), 115-133.
site is determined from the fracture orientation distribution
123
ENTROPY BASED ENSEMBLE NEURAL

NETWORK DESIGN
Zhang Yun (zhangyun@ntu.edu.sg)
Zhao Zhiye (czzhao@ntu.edu.sg)
ABSTRACT: Ensemble neural networks (ENNs) are commonly used networks in many engineering applications due to their better
generalization properties. An ENN usually includes several component networks in its structure, and the component networks commonly
use a single feed-forward network trained with the back-propagation learning rule. In this paper, an ENN, which combines the component
networks by using the entropy theory, is proposed. The entropy based ENN searches the best structure of each component network first,
and employs the entropy as an automating design tool to search the best combining weights of the ENN. An analytical function, namely
Friedman function, is used to assess the accuracy of the proposed ensemble approach. The computational experiment verified that the
proposed entropy based ENN outperforms the simple averaging ENN and the stand alone neural network.
ENSEMBLE NEURAL NETWORKS The resulting maximum entropy probability distribution

corresponds to a distribution which is consistent with the
The artificial neural network (NN) (McCulloch & Pitts 1943) given partial information but has maximum uncertainty or
is a mathematical or computational model for information entropy associated with it.
processing based on the biological neural networks. An
ensemble neural network (ENN) is a collection of a finite To illustrate Jaynes’ principle, we consider a discrete random
number of NNs that are trained for the same task. Usually variable X. To obtain the ‘most objective’ probability
the networks in an ensemble are trained independently distribution of X, the maximum entropy principle can be
and then their predictions are combined (Sollich & Krogh used in the following procedure:
1996). In other words, any one of the component networks
in the ENN could provide a solution or a predictor to the max …(1)
task by itself, but better results might be obtained by an
ENN by combining the solutions that are achieved by the
component networks. The architecture of a typical ENN subject to …(2)
is shown in Figure 1. The two main steps to construct an
ENN are creating component networks and combining these
component networks. and , j=1,2,…,m …(3)
where fj (x) is a given function of x. Using the method of

Lagrange’s multipliers, the resulting distribution is λi (x) =
e–α0 –α1 f1 (xi) –α2 f2 (xi) –…–αm fm (xi), i = 1,2,…,n, where α0, α1, …, αm
are the Lagrangian multipliers which are determined from
the (m+1) constraints in Equations (2) and (3).
DESIGN ENSEMBLE NEURAL NETWORKS

Figure 1. Architecture of a typical ENN. USING ENTROPY CONCEPT
ENTROPY The entropy based ENN can reduce over-fitting in the ENN.
The major steps of the entropy based ENN are shown in
The entropy was introduced in the context of efficiency of Figure 2, which are explained further as follows.
heat engines in early 19th century. According to the second
law of thermodynamics, the entropy never decreases in Creation of the component network can be divided into two
a closed system, and it is a measure of the disorder or steps. The first step is to generate the training data and test
complexity of a system. The maximum entropy formalism data sets, and the second step is to create the component
published by Jaynes (1957) is a fundamental concept in networks. In Step 1, some common ratios of the training
the information theory. The maximum entropy formalism data to the test data will be used in the analyses. The data
is used to determine the probabilities underlying a random used for training each component network are the same.
process from any available statistical data about the process. In Step 2, each component network is created several
times, but the best structure will be used in the ENN.
The criterion to choose the best component network is to
124
Start subject to , Pi > 0 …(6)
where S(P) is the entropy value of the combining weights of

the whole ENN; Pi is the ith component network’s weight
Create an ENN, of the ENN; m is the number of the component networks;
determine the no. of the component μT and μENN are the mean values of the target and the ENN
networks and no. of the hidden nodes output, respectively; σT and σENN are the standard deviations
in each component network. of the target and the ENN output, respectively.
The Lagrangian multipliers can be solved as follows:
For each component NN, randomly Max

run several times, then choose the
best structure, which has the
minimum training-MSE.
…(7)
where λ0, λ1 and λ2 are the Lagrangian multipliers.
Calculate the solution of entropy

equations to determine the Let , i=1,2,…,m, the solution of this problem
weights of each component NN.
is Pi = e–1–λ0 –λ1 • μi–λ2 • σi . Let A = e1+λ0, B = e–λ1, C = e–λ2, and
, the solution of this problem becomes

Use test data to calulate
the test-MSE.
. And Newton’s method is used to solve
end the above equations, so to obtain weights of the component

networks.
Figure 2. Flowchart of entropy based ENN.
A VERIFICATION EXAMPLE
select the one with the smallest training mean-squared-error
(MSE). Since good regression ensemble members must be Friedman #1 is a nonlinear prediction problem which
both accurate and diverse, the training of each component was used by Friedman (1991) in his work on multivariate
network should also have the high accuracy and diversity. adaptive regression splines (MARS). It has 5 independent
Thus, different numbers of hidden nodes would be used predictor variables that are uniform in [0, 1]. The following
in different component networks. The procedure to define Friedman #1 with normally distributed noise (mean 0,
the number of hidden nodes in each component network variance 1) is used to test the entropy based ENN.
is similar to Zhao’s method (Zhao et al. 2008), but the
difference is that choosing the best performance component Y = 10sin(πx1x2) + 20(x3 – 0.5)2 + 10x4 + 5x5 …(8)
networks are based on the component network with the
smallest training MSE. Firstly, 5×5×5×5×5 evenly distributed data along both the
five x-axis and the y-axis are selected from the domain [0, 1]
To use the entropy concept to obtain the unbiased ENN, as the training data for the simulation. Another 4×4×4×4×4
three parts of the problem should be optimized at the same evenly distributed points from the same domain are used
time: to maximize the entropy of the combining weights of as the test data. The maximum training epoch of each
the whole ENN; to minimize the error between the mean component network is set to 30. There are 3125 examples
output of the ENN and the mean target value; to minimize in the training data set, and 1024 examples in the test data
the difference of the standard deviation of the output of set. The 3 kinds of NNs used for the example are: (1) single
the ENN and the standard deviation of the target value. NN, (2) ENN with simple averaging combined method, and
This will be benefiting the whole ENN. The three part (3) ENN with entropy based combined method.
optimizing problem can be formulated as follows:
The number of the input nodes is 5, and the number of the
Problem max …(4) output nodes is 1. The single NNs use 4, 6, 8, 10 hidden
nodes in their hidden layer, respectively. The single NNs
are trained 4 times randomly to find the best results for
min …(5)
125
comparison. In the other two ENNs, there are 4 component CONCLUSIONS

networks. The numbers of hidden nodes in the component
networks are 4, 6, 8 and 10, respectively. After choosing This paper aims to improve the ENN in two aspects: 1)
the best weight configuration of each component network instead of using component NN directly, a preliminary
from 4 random runs, the output of the ensemble networks is selecting process is used to get the best component
combined with the simple averaging method (noted as Ave- NN; 2) the entropy is used to determine the weights of
ENN) for the simple averaging ENN, and the entropy based the component NNs in the ENN. Using the entropy to
ensemble method (noted as EN-ENN) uses the modified combine these best component networks can improve the
entropy value to determine the networks’ weights. performance of the ENN by balancing the contribution of
each component network.
Table 1 shows the corresponding MSE values of the test
data and the training data with 20 runs during training. The Friedman function is used to verify the performance
From Table 1, it can be observed that the EN-ENN provides of the proposed ENN. From the comparison study, which
the best generalization in terms of the mean value and include the single NN, the simple averaging ENN and the
the S.D. for the 20 runs. The relative small SD for both entropy based weighted ENN, it is found that the proposed
ENNs indicates the main advantage of the ENN, i.e. the entropy based ENN outperforms other methods. These
consistency of the NN simulation. The comparison between results also showed the potential of the proposed ENN to
the actual and predicted test results of EN-ENN is shown be applied to other kinds of problems.
in Figure 3. It is noted that all the data points are within
a narrow band of the 45o line.
REFERENCES
Table 1. Results of twenty runs on Friedman #1 function
with 4 component networks. [1] Friedman, J.H., 1991. “Multivariate adaptive regression
Single Ave-ENN EN-ENN splines”. Ann Statist, 19(1): 1-82.
[2] Jaynes, E.T., 1957. “Information theory and Statistical
Minimum 1.052 1.410 1.285
Mechanics I”. Phys Rev, 106: 620-630.
Test-MSE Mean 2.889 1.948 1.862 [3] McCulloch, W.S. and Pitts, W., 1943. “A logical calculus
S.D. 1.679 0.455 0.437 of the ideas immanent in nervous activity”. Bulletin of
Minimum 0.041 0.481 0.181 Mathematical Biophysics, 5: 115-133.
[4] Sollich, P. and Krogh, A., 1996. “Learning with ensembles:
Train-MSE Mean 0.357 0.963 0.584
How over-fitting can be useful” in: Touretzky, D.S., Mozer,
S.D. 0.383 0.358 0.296 M.C., Hasselmo, M.E. (Eds.), Advances in Neural Information
Processing Systems 8, Denver, CO, MIT press, Cambridge,
30 MA, p. 190-196.
[5] Zhao, Z.Y., Zhang, Y. and Liao, H.J., 2008. “Design of
25
ensemble neural network using the Akaike information
criterion”. Eng Appl Artif Intel, 21: 1182-1188.
20
Predicted Value
15
10
5
0
0 5 10 15 20 25 30
Actual Value
Figure 3. Comparison between the actual and predicted
Friedman #1 function test results of EN-ENN.
126
BURST STRENGTH ESTIMATION OF A

CRACKED COMPRESSED NATURAL GAS
(CNG) TANK CYLINDER
Lie Seng Tjhen (cstlie@ntu.edu.sg)
Zhang Baofeng (bfzhang@ntu.edu.sg)
ABSTRACT: In this study, the fracture behavior of a 30-liter cracked compressed natural gas (CNG) cylinder was investigated numerically.
A series of experimental tests was conducted to study the material property of a typical steel CNG cylinder as well as the true profile of
the cylindrical structure. Then, fracture assessment of the cracked steel cylinder was carried out to predict its burst pressure. The approach
adopted in the present analysis was based on the Failure Assessment Diagram (FAD) given by BS 7910 and API 579. Finally, a comparison
was made between the theoretical burst pressure and the ones predicted by both codes of practice.
INTRODUCTION
Fracture failure
1.0
In order to determine the critical crack size with which the Applied load or crack size
crack becomes unstable and causes the CNG cylinder to Failure assessment curve
fail, the damaged structure should be assessed according to Unsafe Plastic collapse
the knowledge of the service stresses and the knowledge Loading path
rdrpVr
Critical point failure

of the fracture properties of the material. The fitness for
service is an important procedure used to indicate the right
level of material and fabrication quality for application
Safe
with regard to the risks and consequences of failure [1]. Service point
Fitness for service assessment procedures for evaluating
crack-like flaws in pressure vessels are based on the failure
assessment diagram (FAD) method, which has evolved as
the most widely accepted methodology for the analysis of Lr =Applied load/plastic collapse load 1.0
components containing a crack-like flaw. Figure 1. Failure assessment diagram (FAD) for
flawed structures.
This study described the structural integrity of a 30-liter
CNG cylinder having a wide range of internal surface cracks.
The load ratio Lr for pressure vessels is provided as
The approach adopted in present analysis was based on the
FAD given by BS 7910 [2] and API 579 [3]. The standard
…(2)
Level 2A FAD curve, which is identical to the Level 2
assessment in API 579, was used to estimate the burst
strength of a typical cracked CNG cylinder specimen. where σref is the reference stress and σy is the yield strength
of the material.
FAILURE ASSESSMENT DIAGRAM (FAD) It is essential to recognize that the Kr parameter uses the
linear elastic stress intensity factor with no allowance for
the effect of plasticity on the crack tip driving force. As
The usage of the FAD for the assessment of flawed structure

is illustrated in Figure 1. This method adopts the assessment Lr increases, plasticity also increases the effective crack
curve which uses the ratio of the stress intensity factor tip driving force [4].
to the fracture toughness, Kr, as the vertical axis and the
ratio of the applied load to the plastic collapse load, Lr,
as the abscissa axis. If the service point falls inside the CNG CYLINDER SPECIMEN
assessment curve, the structure is considered safe, otherwise,
the structure is deemed unsafe. Cylinder geometry
In the standards, the general FAD Level 2A assessment A 30-liter steel cylinder was cut open to reveal the internal
curve is given by structure as shown in Figure 2. The measured dimensions
are external diameter D = 230 mm, side wall thickness t
Kr = (1 – 0.14L2r)[0.3 + 0.7 exp(–0.65L6r)] …(1) = 6.4 mm and overall length l = 670 mm.
127
Table 2. Material properties.

Yield Tensile Young’s Critical Charpy Estimated
stress strength modulus CTOD V-notch Kmat
σy (MPa) σu (MPa) (GPa) δIC (mm) Energy (N/mm3/2)
(Joules)
868.8 968.6 204.5 0.07 101 3464
Table 3a. Assessment results of burst strength (BS 7910).
Figure 2. 3D model of the 30-liter CNG cylinder. BS 7910 (2005)

σh,u (δ route) σh,u (K route)
a c (MPa) (MPa)
Table 1. Principal dimensions of the surface cracks.
deepest crack tip deepest crack tip
Crack depth a Half crack length c
a/t 1.92 9.60 695.33 819.40 681.97 807.72
(mm) (mm)
1.92 3.20 778.92 803.04 766.78 791.19
1.28 6.40 3.20 2.13 1.60 1.28 0.2
1.92 1.92 830.48 807.97 818.61 796.10
1.92 9.60 4.80 3.20 2.40 1.92 0.3
3.20 16.00 514.05 633.22 498.62 622.78
2.56 12.80 6.40 4.27 3.20 2.56 0.4
3.20 5.33 698.20 714.93 684.64 701.97
3.20 16.00 8.00 5.33 4.00 3.20 0.5
3.20 3.20 771.74 739.23 759.65 726.34
3.84 19.20 9.60 6.40 4.80 3.84 0.6
4.48 22.40 293.43 388.11 280.42 379.04
4.48 22.40 11.20 7.47 5.60 4.48 0.7
4.48 7.47 586.81 590.07 572.48 575.89
5.12 25.60 12.80 8.53 6.40 5.12 0.8 4.48 4.48 706.56 658.11 694.04 643.73
For a high pressure compressed natural gas (CNG) tank Table 3b. Assessment results of burst strength (API RP579).
cylinder, a surface crack is often initiated from the internal
API RP579 (2007)
wall thickness and rapidly grows into an approximately
semi-elliptical in shape [5]. Thus, a semi-elliptical internal σh,u (K route)
surface crack was assumed at the side wall of the specimen a c (MPa)
as shown in Figure 3. The assumed surface cracks covers deepest crack tip
a wide range of crack profiles as listed in Table 1, where 1.92 9.60 756.46 913.30
the ratio of crack depth to wall thickness a/t ranged from 1.92 3.20 871.18 879.30
0.2 to 0.8 with an interval of 0.1; the ratio of crack depth
1.92 1.92 930.98 889.40
to crack length a/c ranged from 0.2 to 1 with an interval
3.20 16.00 597.58 762.66
of 0.2.
3.20 5.33 784.40 777.98
3.20 3.20 863.43 805.41
4.48 22.40 464.12 611.25
4.48 7.47 707.55 679.06
4.48 4.48 811.49 729.554
FAILURE ASSESSMENT OF CRACKED CNG

CYLINDER
The safety of this CNG cylinder containing internal surface

crack under a working pressure was assessed according to

Figure 3. Mesh of CNG cylinder model with a surface crack. BS 7910 [2] Level 2A and API 579 [3] Level 2 respectively.
Both stress intensity factor, K, and CTOD, δ, assessment
Material properties routes were employed in the evaluation of the burst strength
of the cracked CNG cylinder. The residual stresses were
The material properties of the CNG cylinder were obtained set as zero because it was a seamless type CNG tank
through tensile coupon test, CTOD test and Charpy impact cylinder, i.e. there were no continuous longitudinal girt
test, and they are summarized and tabulated in Table 2. welds. The selected assessment results of the burst strength
are tabulated in Table 3.
128
CONCLUSIONS REFERENCES
In this study, a 30-liter CNG cylinder with a wide range [1] Wells, A.A., 1981. “The meaning of fitness for purpose and
of semi-elliptical internal surface cracks was assessed to the concept of defect tolerance”. International Conference
predict its burst strength. of Fitness for Purpose Validation of Welded Constructions.
London, UK: The Welding Institute, Paper 33.
Based on the tensile strength and fracture properties [2] BS 7910, 2005. “Guide on Methods for Assessing the
obtained, the prediction was performed using BS 7910 [2] Acceptability of Flaws in Fusion Welded Structures”. British
Standards Institution, London, UK.
and API 579 [3] approaches respectively. Fracture toughness
in terms of critical CTOD value obtained from the CTOD [3] API RP579-1/ASME FFS-1, 2007. “Fitness-for-Service”.
American Society of Mechanical Engineers, New York,
test and critical stress intensity factor KIC estimated from the
USA.
Charpy impact test data were applied in the prediction.
[4] Ainsworth, R.A., 1984. “The assessment of defects in
structures of strain hardening materials”. Engineering Fracture
From the prediction results for the cracked CNG cylinder, Mechanics, 19(4), 633–642.
it can be concluded that the fracture toughness estimated
[5] Lin, X.B. and Smith, R.A., 1998. “Fatigue Growth Prediction
from Charpy impact test data can be used in the fracture
of Internal Surface Cracks in Pressure Vessels”. Journal of
strength prediction for the cracked CNG cylinder. BS 7910 Pressure Vessel Technology, 120(1), 17-23.
[2] approach produces more conservative predictions of
the fracture strength compared to that the API 579 [3]
standard. When the ratio a/c increases, the weakest point
of the crack is expected to shift from the deepest point to
the crack tip.
129
EXPERIMENTAL STUDY AND NUMERICAL

MODELING OF STRESS CONCENTRATION
FACTOR OF HIGH STRENGTH STEEL
PLATE-TO-PLATE Y JOINTS
Chiew Sing Ping (cspchiew@ntu.edu.sg)
Jiang Jin (jian0048@e.ntu.edu.sg)
Yu Yi (yuyi@.ntu.edu.sg)
ABSTRACT: In this study, the stress concentration factor (SCF) distribution of a set of plate-to-plate Y joints made from high strength
steel plates (with yield stress equal to 690MPa) under the action of static axial tensile loading applied on the brace were studied. The
detailed SCF distributions were analyzed via finite element modelling and empirically tested using small scale specimens (with plate
dimensions of 450mm×150mm). Two groups of specimens with different welding procedures are studied: The first group consists of
joints with welding completed at ambient temperature while the second group consists of joints with welding completed at a per-heated
temperature of 100°C. Comparison study is then carried out for the stress concentration factor of high strength steel joint in different
fabrication environments and geometries.
INTRODUCTION
Currently, most steel structures are made of mild steel for

its satisfactory mechanical property and availability. In the
existing codes and standards, mild steel is well specified
for application. However, there has been an increasing
interest in the use of high strength steels, which generally Eqn. 1
have yield strengths larger than 460MPa, for recognizing
the benefits from an increase in the strength to weight ratio
and savings in the cost of materials. This is particularly
applicable to offshore structures to reduce weight which can
lead to achievement of considerable saving in supporting
substructures. Compared with mild steel, high strength steel Figure 2 illustrates geometry profile for the joints whose
has merits in economy, architecture and safety. angle is ranging from 60° to 90. In this case, the end of
attachment plate in the welding part is cut into shape with
On the other hand, fatigue is one of the major problems 30° to satisfy welding requirement. Assuming the coordinate
causing the degradation of offshore structures in long term of point H and G are (l1/2, t1) and (l1/2, t1+R) respectively,
integrity. Therefore, stress concentration factor, which is the coordinate of C, D, E, and F can be expressed as:
the ratio between the hot spot stress, caused by structural
discontinuities or welding, and the nominal member stress,
is of significant meanings.
GEOMETRICAL AND NUMERICAL

MODELLINGS
A systematic geometrical modelling procedure for a general

high strength steel plate-to-plate Y joint without cracks was
proposed in this study. The global coordinate system (x-y)
Eqn. 2
of the joint is defined with the original O locates at one end
point of the plate. The geometry profile for the 45°joints
angle is shown in Figure 1. Assuming that the coordinate
of point C is ((l1+t2)/2, t1), the coordinate of point D can
be written as ((l1+t2)/2, t1+R). Furthermore, the coordinates
of points E, F and G can be expressed as:
130
During numerical modelling using ABAQUS, by using EXPERIMENTAL INVESTIGATION

the importing modulus in ABAQUS and using variable
values of l1, l2 t1, t2 and θ, it was easy to change the In the present experimental investigation, a number of
plate size and welding parameters. Load was applied on plate-to-plate T/Y joints, made of high strength steel with
the surface of EF, as shown in Figure 1 and Figure 2, by minimum yielding stress of 690MPa, were fabricated by
a point-surface coupling. Five monitoring points, a, b, c, welding. This high strength steel, RQT701, which was
d and e, whose z coordinates are 15mm, 45mm, 75mm, supplied by Corus Group, was quenched and tempered
105mm and 135mm respectively, were set for results structural steel with improved forming and welding
analysis in each case (Figure 3) after the modelling were performance by substituting some alloying element with
completed. Linear interpolation was used to get the stress carbon. In the process of welding, great precautions were
in monitoring points after the hot spot stresses at element needed to ensure that welding qualification is satisfactory.
nodes were obtained. Electrodes and fluxes with very low hydrogen content must
be used in order to prevent hydrogen cracking. Hence, an
F
y ultra low hydrogen and moisture resistant type covered
t2
E electrode for 690MPa high tensile strength steel for low
temperature service, LB-70L, which is equivalent to the
l2
class ASME/AWS A5.5 E10016-G and supplied by Kobelco

of Japan, was employed [1]. The welding procedure was
carried out according to the AWS D1.1 2008 [2]. Other
ș standards also were referenced [3-4]. Table 1 gives the
R
H G D B mechanical properties of the RQT701 plate and LB-70L
C x t1 electrode.
tw
o l A
l1 Table 1. Mechanical properties of RQT701 steel plate

and LB-70L electrode.
Figure 1. Welding Profile for 45° joints.
Items Minimum Tensile Minimum
Yield Strength Strength Average
t2 (MPa) (MPa) Impact Energy
F
E RQT701 690 790~930 27J@ -45°C
y
LB-70L 685 755 108J@-60°C
l2
Two series of specimens were included to compare the

influence of welding condition on stress concentration factor
distribution near the weld toe. One group was fabricated in
ambient temperature while the other group was pre-heated
R D ș to 100°C before welding. There were 6 different geometries,
G B
I
H C x t1 consisting of 3 different parent steel plates thicknesses and
o 2 welding connection angles of each group were employed
l A
l1 to explore the variation of the residual stress near the weld

toe. During the welding, full penetration welding for tubular
Figure 2. Welding Profile for 60° joints. joint was used by following the standard AWS D1.1-2008.
For the 45° joints, the welding profile shown in Figure 1.
Figure 2 shows the welding profile for the 60° joints.
To fix the specimen in the grip of testing machine, a set of

supporting joints, made of mild steel S355 with thickness
of 55mm, was designed. It was designed in such a way that

y
have nearly triple thickness when comparing with specimens
to make sure the failure will turn out in specimens rather
a 5 20 than the supporting joints. The specimen and supporting
z b
d
c x joint are connected by 12 high strength hexagon bolts of
e
grade 10.9HR. In each end of connection, 6 bolts were
positioned in two lines. Figure 4 shows the profile of the
specimen and supporting joint after assembly.
The Instron Model 8506 Dynamic Materials Testing System

Figure 3. Measuring points in the specimen. was introduced when tensile loading was to be applied in
the specimen. With maximum tensile loading capability of
2000KN, it is an advanced multiprocessor-based control
131
console which provides full digital control of a testing

system. It consists of a closed load and four columns frame
with movable crosshead, a hydraulic actuator to apply a
force, gripping mechanisms to hold the mechanical test
specimen, and a load cell to measure the force. The position
of the actuator, under closed loop control by controlling the
hydraulic fluid flowing through a servo-valve supplying the
actuator, is measured by a displacement transducer.
FLA-2 of TML strain gauge, which has only one grid, was
adhered on the surface of the specimen. Six strain gauges
were used and installed in the position as Figure 4 shown. Figure 5. The distribution of SCF in monitoring points
In the direction of specimen length, the strain gauges of 8mm–45° joints.
positioned in two lines, which were away from weld toe
5mm and 20mm respectively, were applied.
Weld Toe 5
C B A 20
C1 B1 A1
25 50 50 25
Figure 6. The distribution of SCF in monitoring points

of 8mm–60° joints.
Figure 4. Plane view of scheme of strain gauges

for static testing.
RESULTS ANALYSIS
Figure 5 illustrates the distributions of SCF both by

modelling and test for the 8mm-45° joints, where line
A denotes the specimen in ambient temperature and line
B refers to the specimen fabricated with preheating. The
modelling result agrees well with test outcomes. In this Figure 7. The distribution of SCF in monitoring points
case, the maximum SCF locates at the middle of plate of 12mm–45° joints.
width. The SCF in joint welded at ambient temperature is
slightly higher than in pre-heating joint. Figure 6 depicts
the modelling and test results for the 8mm-60° joints. In
this case, two SCF values in the middle of plate width
deviate much from modelling results and another two

values from test. However, the values in both two ends
agree well. It is the existence of small notch near weld
toe that make two values in the middle of plate width
deviate from the modelling result and the other two values
derived by test.
Figure 7 shows comparison between modelling and test

results for the 12mm-45° joints. The numerical results
corroborate the test results. However, in Figure 8, where the
distribution of SCF for the 12mm-60° joints is illustrated, Figure 8. The distribution of SCF in monitoring points
the values at both ends appear to diverge. At one end of
132
the plate width, the SCF at joint with welding at ambient CONCLUSIONS
temperature is 0.80 while at the other end point in joint
with welding at pre-heating temperature is 2.12. This Based on numerical analysis and test results, several
phenomenon may be due to the uneven distortion of plate conclusions are obtained as follow:
during welding and the asymmetrical applied stress caused (1) In most cases, the SCF values obtained by modelling
by the tensile machine. However, in the middle of plate have good consistence with test results. The numerical
width, the test results are close to modelling value. Figure analysis by using python through ABAQUS is
9 and Figure 10 give the results for the 16mm-45° joints efficient.
and the 16mm-60° joints, respectively. In these two cases, (2) The notch near the weld toe has significant influence
the modelling results corresponds with test values well. on SCF values and it is necessary to avoid producing
such flaws in process of welding. Additionally, the
distortion due to welding may also be an effect of the
distribution of SCF.
(3) In most of the 45° and 60° high strength steel joints,
the SCF locates in the range from 1.4 to 2.3. Several
aberrant values may turn out in the test due to limitation
of the specimens and test setup.
(4) The welding conditions including ambient temperature
and pre-heating temperature do not exert much influence
on SCF.
Figure 9. The distribution of SCF in monitoring points REFERENCE

[1] AWS. ANSI/AWS A5.5. “Specification for Low-Alloy Steel
Electrodes for Shield Metal Arc Welding”. American Welding
Society, Miami, USA. 2006.
[2] AWS.ANSI/AWS D1.1. “Structural Welding Code-Steel”.
American Welding Society, Miami, USA. 2008.
[3] AS/NZS. “Structural steel welding part 4: Welding of high
strength quenched and tempered steels”. Australia/New
Zealand Standard AS/NZS 1554.4, 2004.
[4] BSI. “Eurocode3 -- Design of steel structures. Part1-12:
Additional rules for the extension of EN 1993 up to steel
grades S700”. British Standards Institute, London, UK.
2007.
Figure 10. The distribution of SCF in monitoring points

133
EXPERIMENTAL TESTS OF DIFFERENT

TYPES OF STEEL BEAM-COLUMN JOINTS
SUBJECTED TO CATENARY ACTION
Yang Bo (yang0206@e.ntu.edu.sg)
ABSTRACT: Several structural collapse incidents indicate that failure usually starts from the beam-column joints when exposed to
abnormal loads, especially for steel and composite structures. If the connections are sufficiently robust and there is adequate axial restraint
from adjoining structures, catenary action forms in the beams and slabs. This gives rise to alternate load paths when affected columns
are severely damaged, resulting in large deformations in the beams and slabs. This paper presents experimental results of steel beam-
column joints subjected to catenary action. Two groups of connections (viz. simple and semi-rigid) were studied under column removal
scenarios. Seven experimental tests were conducted. The experimental results demonstrated that web cleat, flush end plate and top and
seat with web angles connections had better performance under extreme loading conditions compared with other types of joints.
INTRODUCTION
The alternate load path method, an important design

approach to mitigate progressive collapse, has been included
by a number of design codes including GSA [1] and DOD
[2]. It is an approach that allows local failure to occur when
subjected to an extreme load, but seeks to provide alternate
load paths so that the initial damage can be contained and
major collapse can be averted. A typical example is shown
in Figure 1 under the scenario when an interior column has Figure 1. Typical example of alternate load path.
been removed by a blast and an alternate load path can
take place through adjacent structural assemblage including and development of mechanical models for analysis and
beams, columns and joints. One of the key mechanisms in design purpose. This paper focuses on the experimental
mitigating the spread of “domino” effect is to redistribute tests of different types of steel beam-column joints under
applied loading on damaged members through catenary catenary action.
action. As shown in Figure 1, the term “catenary action”
refers to the ability of beams to resist vertical loads through In total, 7 experimental tests were carried out on different
the formation of a net-like mechanism. types of steel beam-column joints, including simple and
semi-rigid connections. The experimental tests were
It is noteworthy that beam-column joints are critical carried out in the Protective Engineering Laboratory of
elements of any building structures and they usually Nanyang Technological University. In the group of simple
control the extent of catenary action because of the limited connections, web cleat, top and seat angle, top and seat
resistance and rotation capacity of joints. So far, only with web angles (TSWA) (8mm thickness angles) and fin
very limited research works have been conducted on the plate connections were investigated while in the group of
performance of bare steel connections subjected to catenary semi-rigid connections, flush end plate, extended end plate
action. Most of the reported works focused on welded and TSWA (12mm thickness angles) were studied. The
moment connections [3, 4]. However, in Europe, bolted principal aim of this paper is to provide the experimental
steel connections such as fin plate, flush end plate, web results of steel beam-column joint behaviour, including
cleat and extended end plate, are very popular and the failure modes, development of forces and deflections in
evaluation of these kinds of joints subjected to catenary the beams.
action is important and timely.
A structure research group at Nanyang Technological TEST SET-UP AND SPECIMENS

University, Singapore, is conducting a research programme
to investigate the stiffness, strength and ductility of bolted Test set-up
steel connections subjected to catenary action under the
column-removal scenario. This project involves a series of
The beam-column joint considered for experimental tests
experimental tests on conventional bolted steel connections,
was located above the storey where an internal perimeter
finite element (FE) investigation of connection behaviour,
column had been removed. To simplify the tests, an
134
inflection point was assumed at the middle of the beam

span. Thus, only half of the beam span was used with pin
conditions, as shown in Figure 2. This simplification was
admissible since the focus was on the central connection
above the removed column. With the possible exception
of the floor immediately above the blast, the simplified
specimen was representative of any other upper floors
above the zone of damage since all the joints experienced
a downward rigid body displacement.
(a) Aerial view
Figure 2. Prototype beam-column joint.
The test set-up is shown in Figure 3. Compared with other (b) Front view
joint tests under normal loading conditions, additional Figure 3. The test set-up.
horizontal restraints were provided by an A-frame and
a strong reaction wall to consider the restraint from
surrounding structural elements. In order to consider the TEST RESULTS
rotational restraint from the continuous column to beam-
column joints, a rotational restraint system was adopted, Simple connections
as shown in Fig. 3. In addition, the beams were restrained
from lateral movement. A displacement-controlled point load Before the test was started, the specimen was held at the
was applied to the middle column using an actuator. horizontal position. A vertical load was then applied to
the middle column gradually until fracture occurred in the
connection part.
Test specimens
In the test of web cleat, the specimen reached the largest
In total, seven tests were carried out. Table 1 summaries loading point with fracture of the web cleat close to the
the test specimens. In all these seven tests, M20 8.8 bolts heel, shown in Figure 4.
were used.
Table 1. Summary of the test specimens.
Connection type Beam End

section plate/angle
Web cleat 305×165×40 L90×8
UB S355 S275
Top and seat 305×165×40 L90×8
Simple angle UB S355 S275
connections Fin plate 305×165×40 100×8
UB S355 S275
TSWA 305×165×40 L90×8
UB S355 S275
Extended end 254×146×37 200×12
Figure 4. Failure mode of web cleat.
plate UB S355 S275
Semi-rigid Extended end 254×146×37 200×12
connections plate UB S355 S275 A ductile failure mode was observed in this test and due
TSWA 254×146×37 L150×100×12 to the high rotation capacity, catenary action could develop
UB S355 S275 well.
135
In the test of the top and seat connection, high flexural

action was observed and the failure mode was similar with
the test of web cleat, as shown in Figure 5. However, due
to this high flexural action, catenary action could not form
until the bottom angle fractured.
Figure 6. Failure mode of top and seat with web angles.
Figure 5. Failure mode of top and seat.
In the test of TSWA, the same angle fracture phenomenon

was observed, as shown in Figure 6.
In the test of the fin plate connection, bolts fractured in

Figure 7. Failure mode of fin plate.
shear were observed, as shown in Figure 7.
Figure 8 shows the load-displacement curves for different

types of connections tested in this group. The results from
the experimental tests demonstrated that simple beam-
column joints could provide fairly significant gravity load
carrying capacity under column removal scenarios. The
achievement of these capacities lies with the ability of these
types of connections to accomodate large rotations. If large
rotations can be reached, axial tension stiffness (catenary
action) will overcome moment stiffness as the dominating
mechanism to resisting gravity loading.
The web cleat and TSWA connections exhibited good axial

tension and rotation capacities to develop catenary action. In
contrast, the fin plate and top and seat connections exhibited
only very limited axial tension and rotation capacities.
Figure 8. Load-displacement curves for different types

of connections (simple connections).
Semi-rigid connections
The same loading procedure was used in this group of

tests. Figures 9, 10 and 11 show the failure modes of these
three semi-rigid connection tests.
Figure 6. Failure mode of top and seat with web angles.

Figure 12 shows the load-displacement curves for different
types of connections tested in this group. The results from
the experimental tests demonstrated that flush end plate and
TSWA connections could perform well to develop catenary
action under column removal scenarios due to their large
rotation capacities.
136
CONCLUSIONS
In this study, experimental tests were conducted to

investigate the behaviour of bolted steel beam-column
joints subjected to catenary action. Simple and semi-rigid
connections were studied. The test results demonstrated
the ductility and load capacity of the six connection types
in catenary mode. Among the simple connections in this
study, the web cleat connection had the best combination
of desirable features: ability to develop catenary action
Figure 9. Failure mode of flush end plate. and extremely high ductility (rotational capacity) through
deformation of the web angles. The flush end plate and
TSWA connections had a similar performance with regard to
catenary action in the group of semi-rigid connections.
REFERENCES
[1] General Services Administration (GSA), “Progressive Collapse

Analysis and Design Guidelines for New Federal Office
Buildings and Major Modernization Projects”. 2003.
[2] Department of Defense (DOD), 2005. “Design of Buildings
to Resist Progressive Collapse”. Unified Facilities Criteria,
4-023-03,
[3] European committee for standardization, Eurocode 3: Design
of steel structures—Part 1-8: Design of joints, BS EN 1993-
Figure 10. Failure mode of extended end plate. 1-8: 2005, British Standards Institution, UK, 2005.
[4] Khandelwal, K. and El-Tawil, S., 2007. “Collapse behaviour
of steel special moment resisting frame connections”. Journal
of Structural Engineering-ASCE, 133(5), 646-655.
Figure 11. Failure mode of TSWA.

Figure 12. Load-displacement curves for different types

of connections (semi-rigid connections).
137
MODELING OF PIEZOELECTRIC
ENERGY HARVESTERS
Yang Yaowen (cywyang@ntu.edu.sg)
Tang Lihua (c070073@ntu.edu.sg)
INTRODUCTION Uncoupled circuit model
To date, wireless sensor networks have been widely Regarding the term at the right hand side of Eqn. (1b)
employed for civil structure health monitoring. One as a current source i, this equation actually satisfies the
challenge in their applications is the limited lifespan of Kirchoff’s current law. Under uncoupled assumption, the
the embedded batteries. For the sensor nodes deployed at coupling term χrV in Eqn. (1a) can be dropped. Hence,
inaccessible locations, battery replacement could be tedious the vibration of the PEH ηr and in turn the current source
and expensive. Hence, harvesting ambient vibration energy i in Eqn. (1b) are independent of the electrical load (or
as sensor’s power supply has aroused an intensive research the energy harvesting process). For open circuit condition,
interest in the past few years [1]. Eqn. (1b) is rewritten as
Among all transduction mechanisms, piezoelectric energy …(2)

harvesting is enthusiastically pursued in the literature for
its high power density [2]. Analytical or numerical models which represents a circuit composed of an ideal current
can be established to predict the optimal power from source placed in parallel with the internal capacitance of the
piezoelectric energy harvesters (PEHs) by considering a pure PEH CS. This uncoupled model (Fig. 1(a)) or its equivalent
resistive load. However, for a practical energy storage circuit, model, i.e., an ideal voltage source placed in series with
these approaches are not applicable due to the intrinsic CS (Fig. 1(b)) is usually utilized as the circuit model of
complexity of storage circuit. One effective approach to a PEH [4]. However, such models fail to account for the
address this problem is to derive the circuit model of a resonance frequency shift for different circuit conditions.
PEH, such that the entire system can be simulated. In this In addition, the prediction of optimal load and power under
article, we first derive an accurate circuit model for the uncoupled consumption is inaccurate if electromechanical
fabricated PEH prototype. Subsequently, the entire circuit of coupling is not negligible [5].
the energy harvesting system is established and simulated in
an electronic simulator. Finally, experiments are carried out
on the fabricated PEH. The experiment results, including
the frequency responses of the open circuit voltage, short
circuit current, the power delivered on various resistors as
well as the results of charging various capacitors, validate (a) (b)
the prediction of circuit simulation. Figure 1. Uncoupled circuit models of PEHs.
CIRCUIT MODELING
In fact, if we move χrV in Eqn. (1a) to the righ hand side
as shown in Eqn. (3), χrV obviously plays the role as an
From the beam vibration theory and constitutive equations
additional excitation force imposed on the beam.
of piezoelectricity, the modal electromechanical governing
equations for a cantilevered PEH can be described as [3]
…(3)
…(1a)
When the excitation frequency approaches one resonance
frequency, the magnitude of V will significantly increase
…(1b)
and the term χrV cannot be neglected especially when the
coupling χr is strong. Hence, ηr in Eqn. (3) and the current
where the subscript ( )r represents the r-th vibration mode;
source i in Eqn. (2) are dependent on the electrical load.
ηr, ωr, and χr are the modal coordinate, natural frequency,
damping ratio and modal electromechanical coupling
coefficient, respectively; – früg is the excitation force due
Accurate circuit model
to the base acceleration üg; V and I are the voltage across
the PEH and current output from the PEH, respectively;
To derive the accurate circuit model of PEH, the
and CS is the clamped internal capacitance of the PEH.
electromechanical coupling term χrV should remain in
138
Eqn. (1a). The derivation of the circuit model starts from

this equation.
Analogizing the displacement ηr as charge qr, the excitation

force – früg as a voltage source vr, and the coefficients, 1,
2 ωr, ωr2 and χr of each term at the left hand side as (a) (b)
an inductor Lr, a resistor Rr, the reciprocal of a capacitor
Cr and a transformer with turn ratio Nr, Eqn. (1a) can be Figure 3. r-th branch of the circuit model
when the PEH works in actuator mode.
rewritten as
…(4) When the excitation frequency approaches r-th natural

frequency, Y can be approximated as
Eqn. (4) represents the constitutive equation of an LCR
circuit and satisfies Kirchoff’s voltage law. This circuit is …(7)
actually the r-th branch of the accurate circuit model and
represents the r-th vibration mode of a PEH. Considering
multiple modes, the accurate circuit model of a PEH is in which, Yd and Ymot are termed damped and motional
comprised of an LCR circuit network (Fig. 2). admittance, respectively [3], i.e.,
…(8a)
…(8b)
From (8b), the locus of Ymot in the complex plane near r-th
mode is an approximate circle, i.e.,
…(9)
Using FEA, we can obtain Ymot near each mode. Drawing

the locus of Ymot, we can obtain the parameters Lmr, Rmr,
Figure 2. Accurate multi-mode circuit model of a PEH. Cmr and Nmr by
For a PEH with a regular profile, the parameters Lr, Rr,

Cr, Nr and vr can be easily determined by analogizing Eqns
(1a) and (4) after analytical modal analysis. However, for
a PEH with an irregular and complicated profile, these
parameters should be identified by finite element analysis
(FEA) [3]. The procedure of this approach will be briefly
introduced below.
…(10)
For parameter Lr, Rr, Cr, Nr
Finally, by Eqn. (6), Lr, Rr, Cr and Nr can be determined.
These parameters should be derived from the admittance
of the PEH. The admittance of the PEH is defined by
For parameter vr
…(5)
The last undetermined parameter is the magnitude of the
In FEA, applying the voltage on the PEH, we obtain the voltage source vr for r-th mode. Under a unit excitation
charge or current response, and subsequently the admittance acceleration, vr is related to the charge response exactly at
Y by applying Eqn. (5). The PEH in this case works in r-th natural frequency by the following equation [3],
actuator mode. Hence, the r-th branch of circuit model
(Fig. (3a)) can be converted to Fig. (3b) according to the …(11)
properties of an ideal transformer, i.e.,
Hence, after we obtain the charge response Q(jωr) by FEA,
…(6) vr can be determined by Eqn. (11).
139
SIMULATION AND VALIDATION resonance frequencies as observed in experiment, which

is one prominent characteristic of the electromechanical
Fig. 4 shows the prototype of a fabricated PEH and the coupling effect. Additionally, the magnitudes of Voc and Isc
experimental apparatus. Two piezoelectric transducers, from simulation agree well with those from the experiment.
electrically connected in parallel, are bonded on each Besides, it should be mentioned that the frequency response
side of an aluminum substrate and a steel proof mass is curves in experiment are slightly bent to left (softening
attached at the beam tip. The beam is clamped on a shaker. stiffness), which results from the unavoidable imperfect
An accelerometer is attached on the shaker to monitor the clamping condition [1].
acceleration of the base excitation. Table 1 lists all the
parameters of the PEH prototype. These parameters are
required in FEA to derive the parameters of the circuit
model of the PEH.
Figure 5. Overall diagram of the energy harvesting system.
Figure 4. PEH prototype and experimental apparatus.
By applying the approach introduced in the previous section,

we derive the circuit model of the fabricated PEH, as
shown in the dashed box in Fig. 5. Two vibration modes
are considered to increase the accuracy. Fig. 5 also shows
the standard energy storage circuit (full-wave rectifier D + (a)
storage capacitor Cstorage). The overall circuit of the system
is established in the SPICE software (NI Multisim 10.0
student version). In the following subsections, we report
the simulation that was conducted and compared with the
experimental results.
Open circuit and short circuit responses
The frequency responses of Voc and Isc from simulation

and experiment in the range dominated by the 1st mode
are compared in Fig. 6. The responses are all normalized
by the root mean square (RMS) input acceleration for (b)
fair comparison. It is noted that the derived model can Figure 6. Open circuit volt age and short circuit current
predict the shift between the open circuit and short circuit responses: (a) simulation and (b) experiment.
Table 1. Properties of the PEH.

Item Piezoelectric Transducer Epoxy Layer Substrate Proof Mass
Dimensions (mm×mm×mm) 85×28×0.2 (active volume) 85×28×0.1 178×32×1.5 18.5×32×14
Density (kg/m3) 7750 2200 2700 7850
Modulus (GPa) E11=60.98 E=0.1 E=70 E=200
Poisson’s ratio 0.35 0.38 0.35 0.29
Dielectric constant ε33 /ε0=830
S
– – –
Piezoelectric constant (10-12m/V) d31= -185 – – –
α=1.7, β=2.25×10-5
Rayleigh damping coefficients
(calculated by experimentally measured 1st and 2nd damping ratio ζ1=0.011, ζ2=0.0114 )
140
Optimal power delivered on various resistors There is a minor discrepancy in this frequency between
the experiment measurement and simulation prediction due
To evaluate the optimal power and corresponding optimal to the imperfect clamping in the experiment, as mentioned
load, various resistors are directly attached to the PEH. Fig. in previous section.
7 shows the frequency responses of the power for different
resistive loads. Although the power magnitudes from Fig. 8 shows the voltage waveforms across the five
experiment and simulation are not exactly the same, their capacitors in time domain when the energy harvesting
trends with resistance increase are similar. Both experiment process starts. It is noted that small capacitors are quickly
and simulation results show that the maximum delivered charged to saturation, which means that they are preferable
power first increases and then decreases with the resistance. when instant power supply is required. Additionally, good
The optimal resistance is around 40KΩ. Furthermore, the consistency is observed between the experiment and
frequency to achieve the maximum power (or the peak) simulation results. Hence, with the help of the derived
shifts to right when the resistance increases. These results circuit model, it is convenient to predict the performance of
are reasonable since the increase of resistance corresponds a piezoelectric energy harvesting system by circuit modeling
to the circuit condition shifts from short-circuit to open- and simulation, avoiding the tedious experimental work.
circuit. The simulation captures these phenomena as the
experiment, validating the circuit model that we derived.
Figure 8. Simulation and experiment results for the

PEH charging various capacitors.
(a)
CONCLUSIONS
In this article, an accurate circuit model for a PEH has been

derived. The entire circuit of the energy harvesting system
was established and simulated in an electronic simulator.
Experiments were also carried out on the fabricated PEH.
For the open circuit voltage, short circuit current and the
power harvested for various resistors, consistent trends
of these frequency responses were observed in both
experiment and simulation. The experimental results of
(b) charging various storage capacitors further validated the
prediction from the circuit modelling and simulation. Hence,
Figure 7. Power delivered on various resistors:
(a) simulation (b) experiment. the proposed circuit model of PEHs provides an effective
approach to evaluate the performance of a piezoelectric
energy harvesting system.
Charging various storage capacitors
For a pure resistive load, analytical model or finite REFERENCES

element method can also be used to evaluate the system
performance. However, in practice, the AC voltage from [1] Yang, Y.W., Tang, L.H. and Li, H.Y., 2009. Smart Mater.
the PEH should be first rectified to DC voltage and then Struct., 18:115025.
the harvested energy can be consumed by a load or stored [2] Roundy, S., Wright, P.K. and Rabaey, J., 2003. Computer
in a capacitor. Hence, we further conduct the simulation Communications, 26:1131-1144.
and experiment to evaluate the circuit model when the [3] Yang, Y.W. and Tang, L.H., 2009. J. Intell. Mater. Syst. Struct.,
PEH charges various capacitors. 20:2223-2235.
[4] Ottman, G.K., Hofmann, H.F., Bhatt, A.C. and Lesieutre,
A standard energy storage circuit and the overall system G.A., 2002. IEEE Transactions on Power Electronics, 17:669-
diagram are shown in Fig. 5. We consider five different 676.
capacitors. In both experiment and simulation, the PEH [5] Erturk, A. and Inman, D.J., 2008. Smart Mater. Struct.,
is tuned to vibrate at open circuit resonance frequency. 17:065016.
141
NUMERICAL SIMULATION OF STEEL

BOLTED BEAM-COLUMN CONNECTIONS
SUBJECTED TO DYNAMIC LOADING
Liu Chang (Liuc0014@ntu.edu.sg)
Andrew Tyas (a.tyas@sheffield.ac.uk)
Fung Tat Ching (CTCFUNG@ntu.edu.sg)
Tan Kang Hai (CKHTAN@ntu.edu.sg)
ABSTRACT: This study presents the development of a finite element model with the ability to simulate and analyse the response of
steel bolted end-plate connections subjected to extremely rapid rates of applied loading. The effect of material strain-rate sensitivity
was taken into account and a shear failure criterion was adopted to predict the failure of the connections. By comparing the simulation
results with experimental observed response, it is found that the numerical techniques and the material model used in this study can
predict the structural behaviour of connections subjected to dynamic loading reasonably well.
INTRODUCTION testing, such as the energy absorption, the distribution

of stress in each component of the connections and the
Progressive collapse is one of the most under-researched ductility of the connections.
areas in structural engineering, and yet it holds the key
to the survival of a structure after a blast event. It is well
known that progressive collapse is a dynamic process
because it happens in a matter of seconds. The importance
of steel connections to resist progressive collapse has been
emphasized by many studies over the years. Most of them
focused on investigation of catenary action of the steel
and concrete connections under quasi-static loading [1,
2]. Little test data has been published to demonstrate the
performance of steel connections subjected to rapid rates
of loading. Recently, Tyas’s research team has conducted
Figure 1. The general layout of the testing and
a series of experimental works to determine the response the geometric details of the connection (unit mm).
of typical semi-rigid steel beam-column connections when
loaded rapidly to failure. This experimental programme is
still ongoing and only some initial test results are published FINITE ELEMENT ANALYSIS
[3]. The results show that significant differences appeared
between the behaviour of connections under dynamic and Description of the dynamic testing
static loads.
The finite element modelling was based on the experimental
It is acknowledged that the experimental work under work conducted by Tyas[3], in which both dynamic and
dynamic conditions is much more complicated than the static responses of the flexible end plate beam-column
static one. Safety issues have to be considered for dynamic connetions were investigated. The connection performance
was characterized by a connection tension load vs rotation
testing. Recording the response of structural elements over

relationship. Fig. 1 shows the general layout of testing and
time accurately is also quite difficult. Therefore, numerical
the geometric details of the connection. An 8mm thick
simulations become an attractive option for investigating the
flexible end plate was welded to the beam web by 6mm
behaviour of connections subjected to dynamic loading. welding and bolted to the column flange using M20 Grade
8.8 bolts. In the tests, the dynamic loads were applied from
The main purpose of this study is to develop a finite zero to the peak in 50-60ms.
element model with the ability to simulate and analyse the
response of the steel bolted end-plate connections subjected
to extremely rapid rates of applied loading. The explicit Finite element model
module of commercial finite element software ABAQUS
is used to form the detailed three-dimensional modelling. The testing were modelled using the explicit module of
The simulation results can provide some additional useful commercial finite element code ABAQUS. Information from
information which is difficult to be measured during the previous studies conducted by Krauthammer [4] indicated
142
that, for high rate dynamic loads, ABAQUS has very good been found to be in good agreement with some reported
simulation capabilities. Due to small time durations and test results, these values were selected here to model the
high applied loads required for the current study, the finite material strain rate of the beam and the end plate.
element model was created using 8-noded continuum (brick)
elements with reduced integration (C3D8R). The model
considered contact conditions between the bolts/nuts head
and the end plate/column, the bolt shanks and the holes, the
end plate and the column, respectively. The fixed boundary
condition was applied at the bottom of the beam.
In order to simplify the model and minimise computational

time, the column was modelled as a rigid plate. This is
because the deformation of the column, either globally under
bending or locally at the connection, can be ignored during
the tests. Moreover, only 200mm length of the beam, which
was sufficient to show local deformations of the connection,
rather than the full length, was modelled. An overview of
the finite element model is shown in Fig. 2.
Figure 3. The influence of strain rate sensitivity.
However, there were very limited research the information

on the influence of strain rate on bolts and welds. Therefore,
dynamic increase factors (DIF) of 1.1 developed by
Department of the US Army Corps of Engineers were taken
into account for strain-rate effects of the bolts and welds
when connections were applied at high rate of loading.
Thus the dynamic stress can be determined as:
σd = 1.1σ0 …(2)
Failure criterion
Figure 2. An overview of the finite element model.
As strain-based failure model controls the brittleness/
ductility of the material, a shear failure criterion where
Material model equivalent plastic strain as the failure measure was applied
to simulate the cracking of the connections (elements that
The material model adopted Young’s modulus of 205 GPa are heavily deformed and satisfy the criterion during loading
and Poisson’s ratio of 0.3 for all of the steel components. are automatically deleted from the mesh). Unlike turning
It should be noted that the densities of the beam and the simulations where a predefined fracture line exists, this
column were adjusted to make sure their masses were the damage parameter was activated across the whole analysis.
same in actual test. Plastic stress-stain curves based on It should be noted that this failure model is only suitable
true stress-strain relationship were defined in this study. for high-strain-rate dynamic problems. The research work
Von Mises yield criterion was used to simulate plastic conducted by Hyun Chang Yim et al. [6] suggested the
deformations of the connection. equivalent plastic strain values of base metal and welds
is 0.2 for base metal and 0.1 for welds. These values are
The effects of the material rate sensitivity are very important obtained by comparison with the dissipated energy values
in the high-strain-rate analysis. In this study, the strain rate of welds and base metal in the Charpy-V Notch test.
effects were introduced into the material model by the
widely-used Cowper-Symonds equation as follows:
ANALYSIS RESULTS
σd = σ0 | 1 + ( D ) |
δ q
…(1)
First, the validation of three-dimensional finite element
where D and q are the material parameters, which were models was preceded by comparing their predictions against
determined from experimental results.Various values of static test data using ABAQUS. As described in the test
the constants D and q have been reported in the literature report [3], a static load was applied on the right side of the
for mild steel. Some of the most representative values for column with a pivot, as shown in Fig. 1. In this test, the
the strain rate sensitivity of mild steel are shown in Fig. connection was not highly deformed due to the limitation of
3. Since the values suggested by Marais et al. [5] have the loading system. The relationship between the connection
143
tension load and the connection rotation from experimental Comparisons of the final deformed configuration of the
tests and computer simulations are compared in Fig. 4. As connection and the failure mode between FEA and test
shown in Fig. 4, results from the simulation and that from results are shown in Fig. 6. It can be seen that the numerical
the experiment corresponded very well. It was demonstrated model can successfully predict the failure mode of the
that the FE model including the definition of materials and connection which involves cracking along the fillet weld
the contact conditions is shown to be, reliable for quasi- between the end plate and the beam web. From the test
static connection behaviour. result, we can see that cracking usually starts on one side.
The simulation result can also predict this phenomenon.
Based on the validated FE model, the analysis of the However, it should be noted that the prediction of which
connection under rapidly applied loading was conducted, side of connection crack first is somehow arbitrary.
with strain rate sensitivity and the failure criterion. Fig.
5 shows a comparison between the simulation result and
the test one. Reasonable agreement is also achieved on the
connection load versus rotation curve. The existence of
discrepancy may be due to the simplified loading method
defined as the load reaches the peak in 0.05ms followed
by a constant force during the rest of analysis in the
numerical simulation.
Figure 6. Final deformed connections and the failure mode.
Figure 4. Load vs. Rotation for flexible endplate

connection under static loading.
Figure 7. The development of cracking at T=30ms.
Figure 5. Load vs. Rotation for flexible endplate

connection under rapidly applied loading.
144
The numerical model was validated against the experimental

results conducted at the University of Sheffield, UK (A.
Tyas et al. 2010), in which a series of partial depth end-
plate connections have been tested under both static and
dynamic loading, respectively. The finite element method
allows for further parametric analyses of steel beam-column
connections to provide some comprehensive results which
can be used to form the simplified analytical model for
design, such as the component-based joint model.
REFERENCES
[1] Demonceau, J.F., 2008. “Steel and composite frames: sway

response under conventional loading and development of
membrane effects in beams further to an exceptional action”.
Doctor of Philosophy thesis, Civil and Environmental
Engineering, University of Liège, 2008.
Figure 8. The development of cracking at T=50ms. [2] Yang, B. and Tan, K.H., 2009. “Numerical analyses of beam-
column joints subjected to catenary action under in-plane
loading”. Proceedings of the 7th International Conference
CONCLUSIONS
on Tall Building, Hong Kong.
[3] Tyas, A. et al, 2011. “Dynamic testing of semi-rigid steel
In this study, a finite element numerical model was
beam-column connections”. Proceeding of the COST C26
developed to simulate the response of steel bolted end- International Conference.
plate connections subjected to extremely rapid rates of
[4] Krauthammer, T., Lim, J. and Oh, G.J., 2000. “Findings from
applied loading.
three computer code validations with precision impact test
data”. Proceedings of the 29th Department of Defense Explosive
The explicit module of commercial finite element software Safety Seminar, New Orleans, LA; 2000, p. 18–20.
ABAQUS is used to form the detailed three-dimensional [5] Marais, S.T., Tait, R.B., Cloete, T.J. and Nurick, G.N., 2004.
modelling. The effect of material strain-rate sensitivity “Material testing at high strain rate using the split-Hopkinson
is taken into account using the Cowper and Symonds pressure bar”. Latin American Journal of Solids and Structures
formulation and a dynamic shear failure model was adopted 2004; 1(1): 319–39.
to simulate the development of cracking during the applied [6] Yim, H.C. and Krauthammer, T., 2009. “Load–impulse
loading. characterization for steel connection”. International Journal
of Impact Engineering, 36 (2009), 737–745.
145
EXPERIMENT INVESTIGATION ON
RESIDUAL STRESS DISTRIBUTIONS OF
HIGH STRENGTH STEEL
PLATE-TO-PLATE Y JOINTS
Chiew Sing Ping (cspchiew@ntu.edu.sg)
Jiang Jin (jian0048@e.ntu.edu.sg)
Yu Yi (yuyi@ntu.edu.sg)
ABSTRACT: In this study, an investigation of the residual stress distribution near the weld toe of a set of plate-to-plate Y joins made
from high strength steel with yield stress equal 690MPa are carried out. The hole-drilling method is applied to measure the residual stress
distribution near the weld toe of the joints. Special template and accessories are manufactured to ensure that precise drilling could be
conducted close to the weld toe. The characteristics of the residual stress are discussed and its relationship with the joint geometry and
the welding profile is summarized.
INTRODUCTION near the weld toe. One group is fabricated in ambient

temperature while the other group is pre-heated to 100°C
When comparing with mild steel, high strength steel (HSS) before welding. There are 6 different geometries, consisting
shows relatively poor ductility and therefore the residual of 3 different parent steel plates thicknesses and 2 welding
stress engendered in the process of welding may have connection angles of each group, employed to explore the
significant influence on the fatigue performance of HSS variation of the residual stress near the weld toe. During the
joints. Residual stress not only affects the initiation and welding, full penetration welding for tubular joint is used
onset of the propagation of surface cracks but also changes by following the standard AWS D1.1-2008. The welding
the path of a crack as it grows below the surface. The effect profiles of 45° and 60° joints are shown in Figures 1 and
of residual welding stresses on the performance of welded 2, respectively.
structure is particularly significant when low stresses are
applied. Therefore, for high strength steel structures, it is F
significant to investigate the distribution of residual stress y t2
due to welding in joints. In this research, hole-drilling E
method is applied for checking the residual stress caused
l2
by welding.
SPECIMEN DETAILS ș R
H G D B
In the present experimental investigation, a number of C x t1
tw
o A
plate-to-plate Y joints, made of high strength steel with l
minimum yielding stress of 690MPa, were fabricated l1
by welding. This high strength steel, RQT701, which
Figure 1. Welding Profile for 45° joints.
is supplied by Corus Group, is quenched and tempered
structural steel with improved forming and welding
performance by substitute some alloying element with t2
F
carbon. In the process of welding, greater precautions are
needed to ensure that welding qualification is satisfactory. y
E
Electrodes and fluxes with very low hydrogen content must

be used in order to prevent hydrogen cracking. Hence, an
l2
ultra low hydrogen and moisture resistant type covered

electrode for 690MPa high tensile strength steel for low
temperature service, LB-70L, which is equivalent to the class
ASME/AWS A5.5 E10016-G and supplied by Kobelco of
Japan, was employed[1]. The welding procedure is carried R D ș
G B
I
out according to AWS D1.1 2008[2]. Other standards also H C x t1
are referenced [3-5]. o
l A
l1
Two series of specimens were included to compare the
influence of welding condition on residual stress distribution Figure 2. Welding Profile for 60° joints.
146
TEST SETUP Weld Toe 5
C B A 15
The RS-200 Milling Guide is a high-precision instrument 25
for analyzing residual stress by the hole-drilling method B1
through positioning and drilling of a hole in the center of
a special strain gauge rosette. Its ruggedness and flexibility B2
make it equally suitable for laboratory or field application.
Since positioning precision of the milling guide has great 25 50 50 25
influence on the accuracy of measurement, the RS-200
milling guide with a microscope installed was applied
and secured to samples with quick-setting and frangible
adhesive to bond its foot pads. The milling cutter was
guided carefully to make the cutter progress in a straight
line devoid of side pressure on the hole and friction at the
non-cutting edge. A high-speed air turbine was employed
to form good hole shape and adaptability to incremental Figure 5. Plane view of scheme of strain gauges.
drilling as shown in Figure 3. To make the measured points
close to the weld toe of the joint, a special supporting set 100°C 25°C
was designed.
Figure 3. RS-200 High speed drilling setup.

Figure 6. Comparison of stresses in points A, B and C in y
direction of specimens with 45° and 12mm.
STRAIN GAUGE LOCATIONS
In the measurement of residual stress, a special type of strain Generally, several steps were followed orderly when the
rosette, FRAS-2, which was designed by TML to facilitate hole-drilling method was applied. Firstly, a special strain
positioning three grids on one side of the measurement point, gauge rosette with three grids was bonded at the point
was used to measure the released strain of the specimen where the residual stresses were to be determined as shown
during drilling. In the transverse direction (labeled with y in Figure 5 and Figure 6. Then the RS-200 Milling Guide
in Figure 4), the shortest distance between the weld toe was attached to the test part and centered over the rosette.
and the strain rosette is 5mm. The further one is 15mm Afterwards a precision hole was introduced at the center
away from the weld toe while the furthest strain rosette of the rosette and readings of the relaxed strains were
has a spacing of 25mm from weld toe. In the longitudinal
recorded. Finally, residual stresses were computed.
direction (labeled with x in Figure 5), three strain rosettes
were fixed with distance of 25mm, 75mm and 125mm
away from the edge. It is positioned such that grid 1 is
parallel with direction of applied loading. Figure 5 shows CALIBRATION TEST
the plane view of the strain gauges scheme applied to the
specimens. Two calibration coefficients denoted as and were
determined firstly to calculate the stress from relieved

strains. It was accomplished by installing the residual
strain rosette, FRAS-2, on a uniaxially stressed tensile
specimen, whose size is 200mm × 70mm. It was made of
the steel plate used to fabricate the specimens. The plate
z
was oriented in such a way that grid 1 is parallel to the
x
5 15 25 loading direction (y), placing grid 3 along the transverse
50
25 y axis (x) of specimen when this specimen is fixed in grip
25
50
of tensile machine.
Before drilling, a small loading P, which does not beyond

half of the yield stress of the specimen, was applied to
Figure 4. Strain gauges scheme for residual the specimen to develop the desired calibration stress σc
stresses measurement.
147
and the values of ε1’ and ε3’ were recorded. It is required 45° 60° (1) 60° (2)
that the tensile stress be uniform over the cross section of
the specimen. Then the loading was released and the plate
was removed out form tensile machine. A standardized
dimension hole, whose dimension is the same as in objective
joints, was drilled. The specimen was replaced in the
tensile machine and applied the exactly same loading P as
before drilling and recorded another group strains ε1’ and
ε3’. To make the results more reliable, 8 groups reading
are recorded in different depths ranging from 0 to 2mm
with 0.25mm gap.
TEST RESULTS
Figure 8. Comparison of stress in points A, B and C in y
direction of specimen with preheating and 8mm.
It can be seen that, in Figure 6, for the 45° joints, preheating
can effectively reduce the magnitude of residual stress at
points A, B, C. At point C where x coordinate is 125mm, 45° 60° (1) 60° (2)
the residual stress is compressive for preheating case while
it is tensile for welding in ambient. At point B which is
located at the middle of the plate, the residual stress is
also slightly reduced for preheating case. For 60° joints,
the magnitude of residual stress also can be reduced by
preheating as shown in Figure 7.
In Figure 8, it can be seen that, at points A and B, the

magnitude of stress in the specimens of 60° is higher than
45°, and at point C, the stress in the specimen of 45° is
slightly higher. In Figure 9, the magnitude of stress in the
specimens of 60° is also higher than 45° at points B and
C. It seems that in the middle of plate, the magnitude of Figure 9. Comparison of stress in points A, B and C in y
direction with specimen of roomtemp and 12mm.
residual stress in 60° joints is higher than 45°one.
It is illustrated in Figure 10 for stresses at points B, B1, 100°C (1) 100°C (2)
and B2. It can be seen that the magnitude of residual 25°C (1) 25°C (2)
stress becomes smaller when the distance from weld toe
gets further. But this variation is not linear. The stress is
reduced quickly when the distance goes from 5mm to 15
mm. However, the stress rises slightly from point B1 to
B2. But for the specimens with preheating, the variation
seems somewhat scattered.
100°C (1) 100°C (2)

25°C (1) 25°C (2)
Figure 10. Comparison of stresses in points B, B1 and B2

in y direction of specimens with 60° and 8mm.
CONCLUSIONS
(1) Frequently, the maximum tensile residual stress

perpendicular to weld toe appears in the middle of
plate. In the ends of the plate, either compressive
Figure 7. Comparison of stresses in points A, B and C in y residual stress or tensile stress may exist.
direction of specimens with 60° and 12mm. (2) The magnitude of residual stress reduces greatly
from point B to point B1 for most cases. However, it
148
may increase in a smaller magnitude from B1 to B2. (6) In general, the magnitude of residual stress in x
It means that the relationship between the change of direction is much higher than the residual stress in the
residual stress and the distance away from weld toe y direction. In particular, for one joint, the maximum
is non-linear. residual stress in x direction is located in the middle
(3) Preheating can effectively reduce the magnitude of of the plate. The principle stress can be beyond the
residual stress of 45° joints. For 60° joints, preheating yielding stress in some cases.
can also reduce the residual stress in some cases.
However, it also can increase the residual stress.
Therefore, it is pertinent that evenly high-quality REFERENCES
preheating of the steel plate should be applied during
the welding of high strength steel joints. [1]. AWS. ANSI/AWS A5.5. Specification for Low-Alloy Steel
Electrodes for Shield Mtetal Arc Welding. American Welding
(4) Regarding the effect of welding angle for preheating Society, Miami, USA. 2006.
specimens, the maximum residual stress seems higher [2]. AWS.ANSI/AWS D1.1. Structural Welding Code-Steel.
in 60° joints than 45° joints in most cases. In the room American Welding Society, Miami, USA. 2008.
temperature welding specimens, the angle effect is not [3]. AS/NZS. Structural steel welding part 4: Welding of high
quite obvious. strength quenched and tempered steels, Australia/ New Zealand
(5) For the specimens test in this study with plate thickness Standard AS/NZS 1554.4, 2004.
from 8mm to 16mm, the plate thickness only has [4]. BSI. Eurocode3---Design of steel structures. Part1-12:
slight influence on the magnitude of residual stress. Additional rules for the extension of EN 1993 up to steel
However, it should be noted that for thicker plate, it grades S700. British Standards Institute, London, UK. 2007
could be possible that the plate thickness may have [5]. API.Recommended Practice for Planning, Designing and
more significant effect on residual stress. Constructing Fixed Offshore Platforms, API-RP2A. American
Petroleum Institute, Washington, DC, USA. 1993.
149
ROBUSTNESS OF STEEL ANGLE

BEAM-COLUMN JOINTS UNDER
COLUMN REMOVAL SCENARIOS
Yang Bo (yang0206@e.ntu.edu.sg)
ABSTRACT: Following the World Trade Centre disaster, some researchers have identified joint integrity as a key parameter to maintaining
structural integrity under catenary action and have conducted extensive research works. This paper presents experimental results of
steel angle beam-column joints subjected to catenary action. Three types of angle connections were studied under column removal
scenarios. Nine experimental tests were conducted. The experimental results demonstrate the ductility and load capacities of these three
connection types with different angle thicknesses in catenary mode.
INTRODUCTION numerically. Conventional welded moment and side-plate

moment connections were investigated. Demonceau [6]
After the partial collapse of the Ronan Point apartment conducted a substructure experimental test and five beam-
tower in 1968, engineers began to realise the importance column joint tests in order to observe the development
of structural resistance to progressive collapse. Further of catenary action and its effect on the joint behaviour.
research and design efforts have been directed to this The M-N interaction curves of composite joints (under
area, especially after the World Trade Centre disaster on hogging and sagging moments) were included in his
work [6]. Most of the reported works focused on welded
11 September 2001. The alternate load path method, an
moment connections [4, 5]. However, in Europe, bolted
important design approach to mitigate progressive collapse,
steel connections such as fin plate, flush end plate, web
has been included by a number of design codes including cleat and extended end plate, are very popular and the
GSA [1] and DOD [2]. It is an approach that allows local evaluation of these kinds of joints subjected to catenary
failure to occur when subjected to an extreme load, but seeks action is important and timely.
to provide alternate load paths so that the initial damage
can be contained and major collapse can be averted. One A structures research group at Nanyang Technological
of the key mechanisms to mitigate the spread of “domino” University, Singapore, is conducting a research programme
effect is to redistribute applied loading on damaged members to investigate the stiffness, strength and ductility of bolted
through catenary action. The term “catenary action” refers steel connections subjected to catenary action under the
to the ability of beams to resist vertical loads through the column-removal scenario. This paper focuses on the
formation of a net-like mechanism. experimental tests of steel angle beam-column joints under
catenary action.
Following the World Trade Centre disaster, some researchers
have identified joint integrity as a key parameter to In total, 9 experimental tests were carried out on different
maintaining structural integrity under catenary action and types of steel angle beam-column joints. Three types of
connections, including web cleat, top and seat, and top and
have conducted extensive research works. Khandelwal and
seat with web angles (TSWA) connections were investigated
EI-Tawil [3] used structural simulation to investigate a
while three thicknesses (8mm, 10mm and 12mm) of angles
number of key design variables that influence the formation were tested. The principal aim of this paper is to provide
of catenary action in special steel moment resisting frame experimental results of steel angle beam-column joint
sub-assemblages. Welded joints with and without reduced behaviour, including failure modes, development of forces
steel beam sections were considered. Fahim et al. [4] and deflections in the beams.
conducted an experimental and analytical assessment of
the performance of steel beam-column assemblies with
two types of moment-resisting connections similar to the TEST SET-UP AND SPECIMENS
ones investigated by Khandelwal and EI-Tawil [3] under
the column-removal scenario. As a follow-up work, this A detailed description of the test set-up was given in a
paper investigated the behaviour and failure modes of previous paper [7], so only the specimen details are given
typical steel connections subjected to catenary action under here.
the condition of in-plane loading. In 2009, Karns et al. [5]
conducted a test programme consisting of a steel frame In total, nine tests were carried out. Table 1 summarises
subjected to a blast. The behaviour of different beam-column the test specimens. In all these nine tests, M20 8.8 bolts
joints subjected to blast was evaluated experimentally and were used.
150
Table 1. Summary of the test specimens.

Connection Angle Beam End plate/
type thickness section angle
8mm 254×146×37 L90×8
UB S355 S275
10mm 254×146×37 L90×10
Web cleat
UB S355 S275
12mm 254×146×37 L150×100×12
UB S355 S275
8mm 254×146×37 L90×8
UB S355 S275 Figure 2. Failure mode of web cleat (10mm angle thickness).
Top and 10mm 254×146×37 L90×10

seat angle UB S355 S275
12mm 254×146×37 L150×100×12
UB S355 S275
8mm 254×146×37 L90×8
UB S355 S275
10mm 254×146×37 L90×10
TSWA
UB S355 S275
12mm 254×146×37 L150×100×12
UB S355 S275
TEST RESULTS
Web cleat
The failure mode was observed to change with angle

thickness. When the angle thickness is 8mm, the failure
mode is shown in Figure 1. Failure was caused by the
Figure 3. Failure mode of web cleat (12mm angle thickness).
fracture of the web cleat close to the heel. This fracture
happened at a very high rotation. The two cleats had
undergone a significant amount of deformation at that Figure 4 shows the load-displacement curves of web cleat
stage. All the bolts remained undamaged. connections using different angle thicknesses. When the
angle thickness increases, failure mode also changes from
angle fracture to bolt fracture. However, from the load-
displacement curves shown in Figure 4, no significant
change of the load-carrying capacity was observed. This is
due to the failure mode of angle fracture having a higher
deformation capacity.
Figure 1. Failure mode of web cleat (8mm angle thickness).
When the angle thickness is 10mm, the failure mode is

shown in Figure 2. Similar with the previous test, failure
of this test was controlled by the fracture of the web cleat
close to the heel. Nevertheless different from the first test, Figure 4. Load-displacement curves of web cleat connections.
the bottom bolt in the test deformed significantly.
When the angle thickness is 12mm, no web cleat fracture

was observed. The connection failed by the bolt fracture,
as shown in Figure 3.
151
Top and seat angle TSWA
Figure 5 shows the failure modes of the top and seat angle Figure 7 shows the failure modes of TSWA tests. The same
tests. The same failure modes with web cleat tests were failure modes with web cleat tests were observed.
observed.
(a) 8mm thickness (b) 10 mm thickness

(a) 8mm thickness (b) 10 mm thickness
(c) 12mm thickness (c) 12mm thickness
Figure 5. Failure mode of top and seat angle. Figure 7. Failure mode of TSWA.
Figure 6 shows the load-displacement curves of top and Figure 8 shows the load-displacement curves of TSWA
seat angle connections using different angle thicknesses. connections using different angle thicknesses. Higher
Higher flexural action and load-carrying capacity were flexural action and load-carrying capacity were observed
observed when angle thickness increased. However, catenary when angle thickness increases. With regard to the stage
action could not contribute significantly to the load-carrying of large deformation, all these three connections could
capacities even when the angle thickness was increased. develop catenary action well.
Figure 6. Load-displacement curves of top and Figure 8. Load-displacement curves of TSWA.

seat angle connections.
152
CONCLUSIONS [3] Khandelwal, K. and El-Tawil, S., 2007. “Collapse behaviour

of steel special moment resisting frame connections”. Journal
In this study, experimental tests were conducted to of Structural Engineering-ASCE, 133(5), 646-655.
investigate the behaviour of steel angle beam-column joints [4] Fahim, S., Joseph, A.M., Lew, H.S., Robert, S.D. and
subjected to catenary action. Three types of connections Chiarito, V., 2009. “Testing and analysis of steel beam-column
were studied. The test results demonstrated the ductility assemblies under column removal scenarios”. Proceedings of
the 2009 Structures Congress, USA.
and load-carrying capacity of these three connection types
with different thicknesses in catenary mode. Among the [5] Karns, J.E., Houghton, D.L., Hong, J.K. and Kim, J., 2009.
“Behaviour of varied steel frame connection types subjected to
web cleat connections in this study, angle thickness has a
air blast, debris impact, and/or post-blast progressive collapse
limited influence to load-carrying capacities. In the tests of load conditions”. Proceedings of the 2009 Structures Congress,
TSWA connections, the experimental results indicate that USA.
when angle thickness increases, higher flexural action and [6] Demonceau, J.F., 2008. “Steel and Composite Frames: Sway
load-carrying capacity are obtained. Response under Conventional Loading and Development
of Membrane Effects in Beams further to an Exceptional
Action”. Doctor of Philosophy thesis, Civil and Environmental
REFERENCES Engineering, University of Liège.
[7] Yang, B. and Tan, K.H., 2011. “Experimental tests of different
[1] General Services Administration (GSA), 2003. “Progressive types of steel beam-column joints subjected to catenary
Collapse Analysis and Design Guidelines for New Federal action”. CEE Research Bulletin 24, Nanyang Technological
Office Buildings and Major Modernization Projects”. University.
[2] Department of Defense (DOD), 2005. “Design of Buildings
to Resist Progressive Collapse”. Unified Facilities Criteria,
4-023-03.
153
SEISMIC RESPONSES OF REINFORCED

CONCRETE BUILDINGS WITH
WALL-LIKE COLUMNS
Sahil Bansal (sahi0004@e.ntu.edu.sg)
Huang Yin Nan (ynhuang@ntu.edu.sg)
ABSTRACT: The objective of this study is to investigate the structural response of a typical reinforced concrete building in Singapore,
subjected to a scenario earthquake with a mean annual frequency of exceedance of 2% in 50 years. Two-dimensional finite element
models of a sample 10-storey building with complete infill wall and with no infill in the first storey are analyzed. The results of non
linear analysis show that the latter case has a high value of storey drift ratio at the first storey that may result in significant damage to
the building.
INTRODUCTION in Figure 2. According to the plan of Figure 2, geometric

symmetry is apparent.
Singapore is situated in relatively low seismicity region
30
therefore current design procedures for buildings do not
include any specific provision for seismic loading. Most
buildings are lightly reinforced concrete (RC) structures
designed for gravity and wind loads. Among the RC
buildings, a commonly used structural system is the 20
Concrete Stress, f
buildings consisting of wall-like columns, which are used

to minimize the protrusions of columns into otherwise
usable space (Lim et al. 2009).
10
Studies by Megawati and Pan (2002, 2009) show that
medium and high rise buildings in Singapore might be at
risk in case of a Magnitude-9+ earthquake in the Sumatran Esec
E
subduction zone. This study investigates seismic response
of reinforced concrete buildings with wall-like columns 0
0 0.002 0.004 0.006
subjected to ground motions associated with a mean annual Concrete Strain
frequency of exceedance of 2% in 50 years on a soft soil Figure 1. FEDEAS Concrete Material Model.
site in Singapore. Two-dimensional finite element models of
a sample 10-storey building found in general in Singapore
are developed and analyzed. The impact of infill wall on
modeling and structural performance is discussed.
MODELING
For the analysis part, numerical models of the sample Figure 2. First Storey Layout.
building have been created using nonlinear analysis program

OpenSees developed by the Pacific Earthquake Engineering AMBIENT VIBRATION TEST (AVT)
Research (PEER) Center in UC Berkeley. All beam and
column sections are modeled using the fiber element and To validate the numerical building models, the micro-
rigid floor assumption has been made. Concrete and steel vibration response of the sample building subjected to
material models from FEDEAS structure library have wind and ground ambient vibration was measured using
been implemented to represent the non linear behavior of tri-directional velocity-meters sampling at 100 Hz. This
material. Figure 1 presents the compression envelop and test provides information about the modal parameters of
material parameters for the concrete model. Masonry infill a structure, namely, natural frequency, mode shape and
wall has been replaced by an equivalent diagonal strut damping. The technique used in this study to estimate
model proposed by Madan et al. (1997) using two-nodded the modal parameters of the sample building is Frequency
truss elements. General layout of the building is shown Domain Decomposition (Beincker et al. 2001). It consists
154
of decomposition of power spectral density matrix into Table 1. Natural Frequency Comparison.
single degree of freedom systems using singular value Without Infill With Infill AVT
decomposition. The first decomposed singular value
X-Direction
represents contribution of the dominating mode at that
particular frequency and the corresponding singular vector Mode 1(Hz) 0.46 1.95 1.97
is the mode shape. The remaining represents the lower Mode 2(Hz) 1.38 5.61 4.85
modes or the noise component. Y-Direction
Mode 1(Hz) 0.65 2.06 2.17
To characterize the first two translational frequencies, Mode 2(Hz) 1.78 6.09 ND
measurements have been performed on the 2nd, 9th and
10th floors of the sample building. Figure 3 shows the
SVD values plotted against frequency in the X direction RESPONSE-HISTORY ANALYSIS
and the first two modes can be identified at frequencies
of 1.97 Hz and 4.85 Hz. Ground Motions
Megawati and Pan (2002) studied scenario earthquakes

possible for Singapore and proposed uniform hazard
spectra for different return periods. Figure 4 presents the
spectral acceleration of 10 ground motions generated by
Megawati for this study such that their average matches a
uniform hazard spectrum associated with a mean annual
frequency of exceedance of 2% in 50 years on a soft soil
site in Singapore.
Figure 3. Singular Values vs. Frequency.
Table 1 reports the natural frequencies of the sample

building determined using AVT and eignvalue analysis of
the numerical models. To study the impact of infill wall on
frequencies of structures, the frequencies of the numerical
models for the sample building without and will infill wall
are presented in columns 2 and 3 of Table 1, respectively.
A Young’s modulus of 28000 N/mm2 for concrete is used Figure 4. Acceleration Response Spectra.
for eignvalue analysis to determine the frequencies of
the sample building. The value is corresponding to the Response-History Analysis
initial tangent modulus E, rather than Esec, of Figure 1: a
reasonable assumption for low strain values resulting from Nonlinear response-history analyses were performed to
ambient vibration. investigate the structural response of the sample building
subjected to the spectrally-matched acceleration time series
The model without infill wall significantly underestimates of Figure 4. Only the result for one of the time series of
the frequencies of the sample building identified from the Figure 4 is presented herein. Maximum inter-storey drift
AVT data and the frequencies of the model with infill have ratio is chosen to represent the building performance. Some
a good agreement with the results of the AVT data. The cases of buildings, where the first floor is used as common
contribution of infill wall should not be ignored in the space, do not include wall in the first storey. Analysis for
modeling and analysis of the sample building. such cases is done using the same model but by removing
the wall in the first storey. Figure 5 compares the peak
storey drift of the sample building in the X and Y directions
1) with complete infill wall and 2) with no infill in the first
155
storey. Drift ratio is less than 0.5% for the case with infill The response-history analysis for the sample building
wall throughout the height of the building thus implying that with regular infill shows that the maximum drift ratio is
the damage of the sample building is insignificant. For the less than 0.5% and that for the building with an irregular
case without infill wall at storey one, the distribution of the distribution of wall indicates a maximum drift ratio of 3%
drift ratio indicates a soft storey with a drift ratio of 3% at for the first storey in the X direction.
storey one in the X direction of the sample building. This
drift ratio will result in significant damage in the beams, In this study, the beam-column joints of the sample building
columns and walls of that storey. are assumed rigid, which might underestimate the drift ratio
if the demand of joints exceeds their elastic capacity. More
research is needed to study the impact of the nonlinear
behavior of beam-column joints on the performance of
the sample building.
REFERENCES
[1] Beincker, R., Zhang, L. and Andersen, P., 2001. “Model

identification of output-only systems using frequency domain
decomposition.” Smart Material and Structures, 10, 441-
445.
[2] Chopra, A.K. (1995). Dynamics of structures: Theory and
applications to earthquake engineering. Prentice-Hall, N.J.
[3] Lim, C.L., Li, B. and Pan, T.-C., 2009. “Seismic performances
of reinforced concrete frames with wall-like columns.” The
IES Journal Part A: Civil & Structural Engineering, 2(2),
126-142.
[4] Madan, A., Reinhorn, A.M., Mander, J.B., and Valles, R.E.,
(1997). “Modeling of Masonry Infill Panels for Structural
Analysis.” Journal of Structural Engineering, 123(10), 1295-
Figure 5. Storey Drift. 1302.
[5] Mazzoni, S., McKenna, F. and Fenves, G.L., 2005. “OpenSees
command language manual.” University of California, Pacific
CLOSING REMARKS Earthquake Engineering Center, Berkeley (CA).
[6] Megawati, K. and Pan, T-C., 2002. “Prediction of the maximum
Natural period determined using model with infill is in good credible ground motion in Singapore due to a great Sumatran
agreement with the actual natural period of the building subduction earthquake: the worst-case scenario.” Earthquake
identified from AVT data. The results clearly indicate Engng. Struct. Dyn., 31, 1501-1523.
the importance of incorporating infill wall in numerical [7] Megawati, K. and Pan, T.-C., 2009. "Regional seismic hazard
modeling and its effect on the dynamic behavior of buildings posed by the Mentawai segment of the Sumatran megathrust.”
with wall-like columns. Bulletin of the Seismological Society of America, 99(2A),
566-584.
156
CONSISTENCY OF SHEAR-WAVE
VELOCITY STRUCTURES INFERRED
FROM MICROTREMOR OBSERVATIONS
Daniel Lukas Mulyawan Jap (Daniel.Mulyawan@arup.com)
Irana Pantow (IRAN0001@ntu.edu.sg)
Kusnowidjaja Megawati (kusno@ntu.edu.sg)
ABSTRACT: Shear-wave velocity structures at three locations in Singapore are investigated by the means of array measurement of
microtremor. The microtremor data was analyzed using spatial autocorrelation method and frequency-wave number method. The results
of this study indicate that the shear-wave velocity structures inferred from microtremor observation are consistent with those obtained
using crosshole PS logging. The study also shows that the velocity structure obtained from microtremor survey is independent of the
shape of the array and the time of the measurement. Microtremor survey method is not only applicable for investigating soft soil sites,
but it is also usable for weathered rock sites with thin soil deposits. This is a Final Year Project carried out by the first and second
authors under the supervision of the third author.
INTRODUCTION Referring to Figure 1, the basic scheme and the procedure

of microtremor survey method consist of three steps:
Soft soil deposits are mostly found in Central and South- 1. Observation by seismometer network (array) arranged
Eastern parts of Singapore, namely Kallang and East on the ground surface
Coast. These soil deposits are filled predominantly with
2. Estimation of dispersion of the surface wave (Rayleigh
soft marine clay and alluvial sands. This type of soil
wave) as a response to the subsurface structure directly
would amplify seismic ground motions from earthquakes
below the array
in Sumatra. To quantify the site response amplification, it is
necessary to construct a 3D velocity structure beneath this 3. Estimation of subsurface structure causing the
area, which can be achieved through array measurement dispersion by means of inversion
of microtremors (Okada, 2003). Some of the advantages
using microtremor survey method over conventional drilling
method in subsurface structure investigation are lower cost FIELD OBSERVATION DETAILS AND DATA
and less disturbing effect suffered by the residents due to PROCESSING
noise pollution and vibration.
Field observations were conducted in Katong Park (KAT),
Beatty Secondary School (BES) and Nanyang Technological
University (NTU). These three locations were chosen based
STEP 1 on consideration that Katong Park and Beatty Secondary
School lie on Kallang Formation (soft marine clay), while
NTU lies on Jurong Formation (weathered rock) and the
availability of bore log data at these three locations. Table
1 summarizes field observation details.
STEP 2
Field observations were carried out using 7 portable
microtremor recorder systems. Each set of recorder system
consists of a vertical geophone, a digital recorder and a

Global Positioning System (GPS) antenna, which is used
STEP 3 to identify the coordinates of the point of measurement
and to synchronize the clock of all recorders.
Figure 1. Basic Procedure of Microtremor Survey Method.

RESULTS
Several methods can be used to analyze microtremor data,
such as frequency-wave number (f-k) method, spatial Influence of various array configurations on the
autocorrelation method (SPAC), or H/V spectral ratio. In consistency of SPAC method
this research, only SPAC and f-k method were used for
The objective of the investigation is to examine if the phase
data processing.
velocity of the Rayleigh wave and the shear-wave velocity
157
Table 1. Field Observations Summary.

Location Date Array Configuration
Triangular (7.5 m, 30 m, 400 m
1-10-08
and 1200 m)
Triangular (7.5 m, 30 m, 200 m,
2-12-09
400 m, 600 m and 800 m)
Katong Park
(KAT) Various Small Array
21-01-10
Configurations
Triangular (7.5 m, 30 m, 200 m
6-02-10 and 400 m) and Line
(30 m road side and 60 m)
Beatty Triangular (7.5 m, 30 m, 400 m
3-10-08
Secondary and 1200 m)
School Triangular (10 m, 40 m, 200 m
(BES) 19-03-10
and 400 m)
Nanyang
4-02-10 Triangular (7.5 m and 30 m)
Technological
University
19-03-10 Triangular (200 m and 400 m)
(NTU)
of the subsurface layers inferred from the microtremor Figure 3. Accuracy and consistency of various
measurement is dependent on the configuration of the array configurations.
array. Ideally, the shear-wave velocity profile should
be independent from the array configuration. Several Applicability of microtremor method in estimating
array configurations, as shown in Figure 2, were used to subsurface soil profile
investigate the subsurface structure at KAT.
The accuracy and consistency of microtremor survey
method were investigated using several field observations
mentioned previously. The Observations were conducted
using triangular array with increasing sizes to allow for
estimation of deeper soil layers. The records from small-size
array (7.5 m and 30 m) were analyzed using SPAC method
while the records for medium-size (200 m) and large-size
(400 m) array were analyzed using f-k method.
SPAC method estimates the phase velocity based on the

length combinations of the array. Length combination is
the distance between one sensor to another in an array.
Hence, in a seven sensors triangular array, there exist
five length combinations with their corresponding phase
velocity value. Subsequently the average of the phase
velocity values was used for estimating the soil layer. On
Figure 2. Sketch of various array configurations. the other hand, f-k method estimates the phase velocities by
searching for dominant wave within a wavenumber range
specified. Hence, a good range of wavenumber allow for
The shear-wave velocity profiles inferred from microtremor more accurate estimation of phase velocities.
measurements using different array configurations are

presented in Figure 3. This shows that the estimation
of the shear-wave velocity values of the topmost layer Katong Park measurement
obtained from different array configurations are consistent Theoretical soil layers constructed from the dispersion
and agreeable with the value obtained from crosshole PS curve estimated using SPAC and f-k method for the three
logging (reference value). The thickness of the topmost layer measurements in KAT are shown in Figure 4 below. The
ranges from 25 to 31 m among various array configurations, comparison of the theoretical soil layers with conventional
and it is thinner than the 40-m thickness revealed by the bore log data is presented in Figure 5 below. It can be
crosshole PS logging. It should be noted that microtremor seen that microtremor provides a reliable estimation for
survey method is not able to capture the gradual increase subsurface soil structure.
of shear-wave velocity from the depth of 33 to 45 m as
Based on investigations using array of size larger than
revealed by the crosshole PS logging.
600m, it was found that long-period Rayleigh wave did not
158
have enough energy to cause microtremor wave propagation Beatty Secondary School measurement
throughout the large array area. This finding implies that
KAT is a relatively calm area with low level of vibration The theoretical soil layers constructed from BES
caused by long-period oceanic wave. Hence, large array measurement are shown in Figure 6 below. Figure 7 shows
recordings were not used for further analysis of KAT. the comparison of soil profile estimated form microtremor
measurement and soil profile constructed from bore log data.
An observation using large array was also conducted and
it is concluded that similar to KAT, the area is relatively
calm with little long-period microtremors. The results from
both KAT and BES also showed that the recordings obtain
from the several observations conducted yield consistent
subsurface structures.
Figure 4. (from top) Theoretical soil profile estimated from Figure 6. Theoretical soil profile estimated for BES area based
1 October 2008, 2 December 2009, and 6 February 2010 on 3rd October 2008 and 19th March 2010 observations.
measurements at Katong Park.
Figure 7. Comparison of soil profile obtained from bore log

with profiles estimated from microtremor observations
Figure 5. Comparison of soil profile obtained from bore for Beatty Secondary School.
log with profiles estimated from microtremor observations
for Katong Park.
159
Nanyang Technological University (NTU) CONCLUSIONS
This observation was done to assess the feasibility of Shear-wave velocity structures at three locations in
microtremor method applied in hard soil stratum. The Singapore were investigated by the means of array
observation result with the estimated soil profile is shown measurement of microtremor. The microtremor data
in Figure 8 below. The comparison of the estimated soil was analyzed using spatial autocorrelation method and
profile with nearby bore log exploration result is presented frequency-wave number method. The results of this study
in Figure 9. The soil strength is represented by SPT N indicated that the shear-wave velocity structures inferred
values in the bore log data. The N value was then converted from microtremor observation are consistent with those
into shear wave velocity value using formula extracted obtained using crosshole PS logging. The study also showed
from report by BCA (2006). The finding implies that that the velocity structure obtained from microtremor
microtremor method could provide a reliable estimation survey is independent of the shape of the array (line,
for hard soil layers. square, rectangle, and equilateral triangle) and the time of
the measurement. Microtremor survey method is not only
applicable for investigating soft soil sites, but it is also
usable for weathered rock sites with thin soil deposits.
REFERENCES
[1] Building and Construction Authority, 2006. Evaluation of

Site Response in Singapore due to earthquake effects (part
ii). [unpublished].
[2] Brigham, E.O., 1988. The Fast Fourier Transform and Its
Figure 8. Theoretical soil profile estimated for NTU area Application. Prentice Hall.
based on February and March 2010 observations. [3] Capon, J., 1969. High-Resolution Frequency-Wavenumber
Spectrum Analysis. IEEE, (pp. 1408-18).
[4] Chavez-Garcia, F.J., Rodriguez, M. and Stephenson, W.R.,
2006. Subsoil Structure Using SPAC Measurements along a
Line. Bulletin of the Seismological Society of America, Vol.
96, No. 2, 729–736.
[5] Horike, M., 1985. Inversion of phase velocity of long-period
microtremors. J. Phys. Earth. 33, 59-96.
[6] Ohori, M., Nobata, A. and Wakamatsu, K., 2002. A Comparison
of ESAC and FK Methods of Estimating Phase Velocity
Using Arbitrarily Shaped Microtremor Arrays. Bulletin
of the Seismological Society of America, Vol. 92, No. 6,
2323–2332.
[7] Okada, H., 2003. The Microtremor Survey Method. Society
of Exploration Geophysicists.
[8] Toksöz, M.N., 1964. Microseisms and and attempted
application to exploration. Geophysicics, 29, 154-177.
Figure 9. Comparison of soil profile obtained from bore log

with profiles estimated from microtremor observations for NTU.
160
RESEARCH PROJECTS
ONGOING PROJECTS
A partial list of research projects is summarized below. Readers are welcome to email the respective investigators for more
information regarding their work.
PROJECT TITLE PRINCIPAL INVESTIGATOR
Underwater Infrastructure and Underwater City of the Future Chu Jian

cjchu@ntu.edu.sg,
Teng Susanto
csteng@ntu.edus.g,
Tan Soon Keat
ctansk@ntu.edu.sg
Sustainable Urban Waste Management for 2020 Wang Jing-Yuan,

jywang@ntu.edu.sg
Ng Wun Jern
wjng@ntu.edu.sg
Underground Technology and Rock Engineering (UTRE) Phase II Ma Guowei

cgwma@ntu.edu.sg
Aquaporin Based Biomimetic Membranes For Water Reuse and Desalination Anthony Gordon Fane
agfane@ntu.edu.sg
TEC Project. Nanostructures Photocatalyst for Membrane Fouling Control Sun Delai, Darren
ddsun@ntu.edu.sg
Development of a GIS Based System for Earthquake Response Monitoring of Kusnowidjaja Megawati
Buildings in Singapore kusno@ntu.edu.sg
Enhanced Biological and Physical Stabilization in Landfills Ng Wun Jern

wjng@ntu.edu.sg
Improving The Efficiency of Membranes in the Water Industry_Project 176: Tang Chuyang
Novel Technologies for Enhanced Control of Concentration Polarisation and cytang@ntu.edu.sg
Fouling in Reverse Osmosis Membrane Processes
Environmental Technology of Brine and Reject Streams_ Project 174: Anthony Gordon Fane
Brine Processing By Membrane Distillation Crystallization agfane@ntu.edu.sg,
Law Wing-Keung Adrian
cwklaw@ntu.edu.sg
Environmental Technology of Brine and Reject Streams_Project 173: Tang Chuyang

Novel Pressure Retarded Osmosis (PRO) Technology for Cost-Effective cytang@ntu.edu.sg,

and Environmentally Sustainable Desalination Brine Disposal and Law Wing-Keung Adrian
Osmotic Power Harvesting cwklaw@ntu.edu.sg
Novel high energy density vanadium redox flow cell for renewable Tuti Mariana Lim
energy storage tmlim@ntu.edu.sg
Prediction of Explosion Hazards from Earth Covered Magazine Fan Sau Cheong
cfansc@ntu.edu.sg
Nitrogen-Doped TIO2-Activated Carbon (AC) Composite for Adsorptive Lim Teik Thye
Photocatalytic Oxidation-Reduction of Refractory Organic Substances cttlim@ntu.edu.sg
Under Solar Irradiation in Water Purification
161
RESEARCH PROJECTS
Structural Resilience Study of Concrete Precast and Composite Steel Joints Tan Kang Hai
Subject to Missing Column Scenario ckhtan@ntu.edu.sg
Development of Analytical Tools for Progressive Collapse Analysis due to Tan Kang Hai
Terrorist Bombing ckhtan@ntu.edu.sg
Improving The Efficiency of Membranes in the Water Industry_Project 175: Law Wing-Keung, Adrian
Novel Hydrodynamics for Low Pressure Membrane Processes cwklaw@ntu.edu.sg
Sand Accretion Study Tan Soon Keat

ctansk@ntu.edu.sg
Assembling Of Multifunctional TIO2 Nanofiber Membrane For Sun Delai, Darren

Water Treatment (0802-IRIS-06) ddsun@ntu.edu.sg
Integration of Novel Forward Osmosis Membranes and Optimized Anthony Gordon Fane
Bioprocess for Water Reclamation agfane@ntu.edu.sg
Plane wave absorbers for wave power generation Law Wing-Keung, Adrian
cwklaw@ntu.edu.sg
Improving The Efficiency of Membranes in the Water Industry_Project 178: Anthony Gordon Fane
Sensors For Fouling Control in Reverse Osmosis Membrane Processes agfane@ntu.edu.sg
Emerging Organic Contaminants in Catchment Surface Waters of the Chang Wei-Chung

Marina Bay wcchang@ntu.edu.sg
MBR Process Modeling And Optimization: Case Study of Ulu Pandan Jim Chen Chin-Kuang
Water Reclamation Plant With Future Scale-Up Considerations jimchen@ntu.edu.sg
MPA-NTU Joint Collaboration on Maritime Innovation and Tan Soon Keat

Strategic Technology ctansk@ntu.edu.sg
Explore Concept of Membrane Action in Slabs to Reduce Fire Protection Tan Kang Hai
for Beams ckhtan@ntu.edu.sg
Slope Instrumentation for the Study of Rainfall -Induced Slope Failures Harianto Rahardjo
in Singapore chrahardjo@ntu.edu.sg
Conversion of Municipal Plastic Waste into an Innovative Wang Jing-Yuan

Polyhdroxyalkanoate (PHA) Material jywang@ntu.edu.sg
Development of Novel Hollow Fiber Membranes Integrated with Wang Rong

Biological/Biomimetic sorption for CO2 separation from Biogas rwang@ntu.edu.sg
Energy efficiency and indoor air-quality control in air-conditioned buildings Chang Wei-Chung
wcchang@ntu.edu.sg
The Jurong rock cavern project at Banyan Basin, Jurong Island Low Bak Kong
(JTC C05502007) bklow@ntu.edu.sg
Slope repair and technology in Singapore Harianto Rahardjo

chrahardjo@ntu.edu.sg
Integration Development of Flexible DSSC for Commercial Application Sun Delai, Darren
ddsun@ntu.edu.sg
162
RESEARCH PROJECTS
Development and Assembling of High Efficiency Dye Sensitized Solar Cells Sun Delai, Darren
and Water Cleavage-Hydrogen Production Reactor Using Novel Nano ddsun@ntu.edu.sg
Structured Ti)2 Fiber/Tube/Membrane
Biocement-A New Sustainable And Energy Saving Material For Construction Chu Jian
And Waste Treatment cjchu@ntu.edu.sg
Failure Modes and Ultimate Strength of Tubular Joints under Elevated Tan Kang Hai
Temperatures ckhtan@ntu.edu.sg
The Jurong rock cavern project at Banyan Basin, Jurong Island Low Bak Kong
(JTC C05502007)05 bklow@ntu.edu.sg
The Influence of Floor Slabs and Transverse Beams on the Behavior of Li Bing
RC Beam-Column Joints under Loss of Column Scenarios cbli@ntu.edu.sg
Off-line Portable Damage Detection Devices for Compressed Natural Lie Seng Tjhen
Gas (CNG) Cylinders Fitted in Vehicles ctlie@ntu.edu.sg
163
RESEARCH PROJECTS
COMPLETED PROJECTS
Strengthening techniques to increase seismic 1.0 To study suitable location for sensors to be installed
resistance in school building in Hebei & in an estimated number of 6 buildings, where
extensive instrumentation with a minimum of six
Liao Ning, China
tri-axial seismic sensors (or episensor force-balance
Principal Investigator: Li Bing
accelerometers) linked to a data logger shall be done
Report No.: CEE/2010/192
for each building.
The project beneficiaries include 1,000 students who will
2.0 To study suitable location for the sensors to be installed
upon project completion benefit from a safer school building
for the remaining buildings, where a minimum of two
being repaired and strengthened against earthquakes.
tri-axial seismic sensors (or episensor force-balance
Through the training of the trainers, 10 master trainers and
accelerometers) linked to a data logger should be
40 local builders will also benefit.
placed in each building for monitoring purposes.
3.0 To perform computer modeling, structural analyses

Sichuan earthquake - a drop of Hope and to give recommendations on the followings:
Report No.: CEE/2010/193 (a) Propose and use special purpose software to
develop detailed analytical models of 20 buildings
Construction of Mobile Water Purification System (MWPS) and then analyze the buildings using non-linear
which is capable of producing an estimated output of 60 static and dynamic analyses based on the latest
m3/day of drinking water (using reverse osmosis) and 180 as-built structural drawings. The models shall
m3/day for general usage (using ultra filtration). account for material and geometric non-linear
behavior, including strength/stiffness degradation,
ductility of structural elements and joints, second
order effects and also take into consideration
Sichuan earthquake - New School New Hope amplifications of seismic motion due to soil and
structural resonance.
(b) Determine structural responses and capacity curves
Seismic strengthening and repair of school building for the selected buildings under study in terms
against earthquakes using carbon fibre, steel angles and of accelerations and lateral drifts at base, mid
wire mesh height and roof levels of the building and base
shear through the analyses.
(c) Study the seismic performance of high-rise
Sichuan earthquake - New School New Hope buildings in Singapore.
Underground Technology and Rock Engineering
The project beneficiaries include 1,000 students who will (UTRE) Program - Behaviour of Rock Cavern
upon project completion benefit from a safer school building
Under Dynamic Loads
being repaired and strengthened against earthquakes.
Principal Investigator: Zhao Zhiye
Through the training of the trainers, 10 master trainers and
40 local builders will also benefit.
The study includes two parts: the behavior of rock caverns

under dynamic loads and the analysis of drill/blast operation
Consultation Study On The Tremor Monitoring in rock excavation.
System (TMS) For 30 Buildings
Principal Investigator: Kusnowidjaja Megawati (1) Numerical simulations were conducted to investigate
Report No.: CEE/2010/196 the blast loading and blast induced damage into rock
mass and rock cavern based on continuous modelling.
NTU has recently been appointed as Tremor Assessment The fully-coupled method was used to obtain the blast
Specialist (TAS) in a project commissioned by BCA to load and the damage depth of rock mass. The PPV
instrument 30 high-rise buildings in Singapore. In this damage criteria, taking into account of the influence of
project, NTU is partnering with Ryobi Geotechnique Pte RMR, were used as the damage criterion. The damage
Ltd. Dr Zhang Qiwei, the proposed Research Fellow, is to depth into rock mass including the effect of loading
work on this project. The scopes of the work are: density, weight of charge and rock mass properties was
164
RESEARCH PROJECTS
obtained. The effects of rock bolts were also studied Myanmar Cyclone - Wellspring of Hope
based on the continuous model. The effects of stand- Principal Investigator: Li Bing
off distance, rock bolts and joint orientation on the Report No.: CEE/2010/199
dynamic response of rock cavern were investigated
based on discontinuous deformation analysis (DDA). It Providing safe drinking water by extracting ground water
can be concluded that the displacements and velocities (drilling tube-wells) or surface water (pumping river water)
of a certain measure point near rock tunnels under blast complete with appropriate water treatment facilities (sand
loading are highly dependent on the stand-off distance filters with backwash and chlorine disinfectants), storage
from the charge hole, the overburden of the tunnel, tanks, diesel engine and compressors for the Cyclone Nargis
and joints orientation. The rock bolts can slightly affected communities which are served by the Sun Quality
decrease the displacement and velocity of measure Health clinics at 20 locations in the Irrawaddy delta.
points near it while its effects are insignificant.
(2) Rock blasting is a rock excavation technique widely

used in the mining and construction industry due to its
reliability, economy and safety. The goal of blast design
is to attain the expected technical target (advance and
good contour) at an economical cost.
In addition, the vibration caused by rock blasting should

be taken into consideration, and be controlled under a
safe level to ensure the stability of adjacent structures.
An extensive literature review on the various drill/blast
methods was carried out, and the numerical modelling
using both the continuous and discontinuous methods
(LS-DYNA and DDA) was conducted to investigate
the influence by various design parameters. A case
study was included in the report.
Autonomous Verification and Validation for

Simulation Modeling
Principal Investigator: Yang Yaowen
Simulation models have been increasingly used in problem

solving and in decision making. The correctness of a
simulation model is addressed through model verification
and validation, which is a significant factor to determine
the accuracy and confidence level of the simulation
model. Unfortunately, there has been no set of tests that
can be easily applied to determine the correctness of the
model. Furthermore, no algorithm exists to determine
what techniques or procedures to use. The tool of model
validation and verification is thus highly desired in the
simulation industry, especially in the situation without
test data or with incomplete test data. Model validation

approaches to assess the quality of a simulation model
will be explored in this project. An effective model
validation method will be proposed to conduct model
validation with incomplete test data. It will cover data
validity, conceptual model validity, computerized model
verification, and operational validity. A process of model
validation will be proposed and implemented as part of
modelling environment.
165
RESEARCH PROJECTS
PhD THESES
The Occurrence and Molecular of the DDA in the past decades were given. The formulae
Characterization of Enteric Viruses in the of the NDDA for program were also presented as a main
work of this thesis. The analytical solutions for the inertia
Tropical Aquatic Environment
matrix and contact matrices which control the stability of
Candidate: Aw Tiong Gim
the open-close iterations of block kinematics of the NDDA
Report No.: CEE/PhD/2010/211
were provided and discussed. The NDDA is not a simple
couple of the FEM and the DDA but a hybrid of them. It
Enteric viruses have emerged as important causes of
can work at three states: pure FEM, pure DDA and mixed.
major waterborne disease outbreaks in recent years. The
When the system is continuous, the NDDA works in the
occurrence of enteric viruses in aquatic environments
pure FEM mode. When the system is totally discontinuous,
constitutes a major health hazard because of their low
the NDDA works in the pure DDA mode. Otherwise, the
infective dose and resistance to environmental breakdown.
NDDA works in a mixed mode which retains the advantages
Currently, our knowledge on the occurrence of enteric
of both methods.
viruses in tropical water environments is limited. This
thesis describes the first study of the prevalence and
The idea of NDDA is easy and ready to be realized
molecular characterization of waterborne pathogenic viruses
since the FEM and the DDA are both derived from the
in urban water catchments and community wastewaters in
minimization of the total potential energy of the system.
Singapore. Noroviruses were determined to be the most
An FEM code can be easily transformed into a DDA code
prevalent enteric viruses detected in both wastewater and
when the kinematics part is considered. The NDDA can
surface water samples. Molecular characterization revealed
absorb both the advantages of the FEM and the DDA.
the genetic diversity of astroviruses, enteroviruses and
To transform an FEM algorithm into a DDA algorithm,
noroviruses in water environments. The norovirus strains
two steps are necessary: (1) scheme for the fracture of
detected in environmental samples are closely related to
the continuous material; (2) introducing of the inertia and
the concurrently identified clinical strains in Singapore
kinematics matrices.
during gastroenteritis outbreaks between August 2006 to
January 2007. Knowledge on the occurrence of human
Lastly, numerical simulations were performed to show the
enteric viruses in water environments contributes to the
improvement and flexibility of the NDDA over the standard
understanding of the mechanisms of viral transmission
DDA. The simulation of stress wave propagating inside a
and the possible role played by water as a vehicle of
rock bar shows the ability of the NDDA in dealing with
transmission.
wave problems and its fracturing ability when the stress
wave cause breakage. The simulation of rock specimen with
initial cracks under uniaxial forces shows the procedure of
Nodal-based Discontinuous Deformation crack propagation which agrees with empirical findings.
Analysis The simulations of Brazilian disc test series also agree well
Candidate: Bao Huirong with the experimental results. Indeed, the NDDA can be
Report No.: CEE/PhD/2010/212 applied to more engineering problems other than the crack
propagation problems if more mature FEM algorithms are
This thesis presents a nodal-based discontinuous deformation applied to them.
analysis (NDDA) based on the coupling of the discontinuous
deformation analysis (DDA) and the finite element method
(FEM), for modeling blocky systems, especially for Sorption of Oxyanions on Nanocrystalline
simulating crack propagation in rock mass. The NDDA can
provide a more accurate stress and strain distribution in
Mg/Al Layered Double Hydroxides: Sorption
Characteristics, Mechanisms, and Matrix
each block and has a higher computational efficiency than

the standard DDA in dealing with continuum materials. As Interferences
an important ability of the NDDA, the fracturing capability Candidate: Goh Kok Hui
that allows for shear or tensile fracturing of intact blocks Report No.: CEE/PhD/2010/213
has been provided based on the Mohr-Coulomb law. A
computer program called 2D-NDDA was developed to Nanocrystalline Mg/Al layered double hydroxide (LDH)
handle the combination of continuous and discontinuous with nitrate intercalation produced in this study exhibited
large displacement problems, as well as large deformation excellent affinity for polyvalent oxyanions, but comparatively
and failure analysis, under external loads and boundary less affinity for monovalent oxyhalides. Nanocrystalline
conditions. LDH was prepared by a fast coprecipitation with
subsequent hydrothermal treatment method. The synthesized
After a brief introduction of the concept of the standard nanocrystalline LDH possessed mesoporous characteristic
DDA, detailed reviews of the validation and enhancement with a large surface area and comprised nanocrystalline
166
RESEARCH PROJECTS
grains. Interactions of oxyanions (i.e. arsenate, chromate, and length of the suction zone and time on bedload transport
vanadate) and oxyhalide (i.e. bromate) with nanocrystalline rate. Two different types of cohesionless sand particles
LDH were studied through stoichiometric calculations, with diameters = 0.9 mm and 0.48 mm were used in the
nitrate displacement investigations, comprehensive study.
sorption/desorption experiments, and analyses with several
microscopic techniques such as XPS, EXAFS, XRD, FTIR, Three dimensionless groups, viz. Einstein’s parameter
CHNS/O, and EDX. The influences of co-existing species Φ, Shields’ parameter without seepage τ*o and modified
on the sorption of oxyanions by nanocrystalline LDH were densimetric Froude number Ω are chosen to examine how
investigated by conducting experiment in the presence of seepage affects bedload transport rate. Eleven undisturbed
natural organic matter (NOM) and common anions such as flow conditions were tested, five under clear water and
nitrate, silica, sulfate, carbonate, and phosphate. Arsenate six in live-bed condition. No bedforms with significant
sorption performances of LDHs prepared by various height were observed during the experiments. The results
alternative synthetic routes were also explored and compared show that an increase in suction rate causes an increase in
with those of LDHs prepared by conventional routes. shear velocity excess, which is defined as the difference
between the bed and critical shear velocities, leading to an
increase in bedload transport rate. The experimental data
Exhausted Carbon for the Removal of also show that for the same undisturbed flow conditions,
Hydrogen Sulfide and Ammonia the dimensionless bedload transport rate increases linearly
Candidate: Jiang Xia (in semi-logarithmic scale) with increasing suction rates.
Report No.: CEE/PhD/2010/214 The equations for predicting bedload transport rate under
suction are derived empirically; the predicted results using
The main objective of this study is to develop a highly these empirical equations compared well with measured
efficient and economical biofiltration process for the data, with an accuracy of ±20%. Published results from
co-removal of H2S and NH3 using exhausted carbon other researchers are used to compare with results obtained
as packing material. Firstly, the feasibility of re-using and inference drawn from the present study. The similarities
exhausted carbon in biofilters for the removal of H2S was and differences of these studies are highlighted.
confirmed. The removal efficiency of H2S was almost
identical in the biofilters packed with exhausted carbon On the other hand, the experimental results show that
and fresh carbon. Furthermore, a mathematical model was the sediment transport rate decreases with an increase
developed to explore and explain the different mechanisms in injection velocity, but the magnitude of the reduction
of H2S removal in the two biofilters. The profiles of H2S is comparatively smaller than that of the increment with
concentration along the biofilm thickness and carbon radius suction. The equation for predicting bedload transport
were simulated using the proposed model system. The rate with injection is also empirically determined by fitting
ratios of the H2S removal by the mechanism of adsorption the ratio of Einstein’s parameter Φ with injection to that
and biodegradation in the biofilters were also simulated, without as a function of the ratio of modified densimetric
respectively. Thereafter, the effect of substrates acclimation Froude number with injection to that without.
strategy on simultaneous biodegradation of NH3 and H2S
was evaluated. Different biodegradation capacities of The effect of the length of the seepage zone on sediment
NH3 and H2S were observed under different substrates transport rate was investigated experimentally in this
acclimation strategies. Lastly, a horizontal biotrickling filter study. Only suction effects were studied in this part of the
(HBTF) packed with exhausted carbon for the co-removal research. The results show that for the same undisturbed
of H2S and NH3 was investigated. The results demonstrate flow conditions and suction rate, the bed load transport
that it is highly efficient and effective for simultaneous rate, which is represented by Einstein’s dimensionless
biodegradation of H2S and NH3 by the HBTF over 316 parameter Φ, decreases with a reduction in suction zone
days of operation. The long-term high performance of length. A slope modifier is introduced to account for the
the HBTF is attributed to low accumulation of biomass length effect. The empirical equation for predicting the
slope modifier is obtained in terms of the Shields’ parameter
and products, stable carbon characteristics and microbial

communities. without seepage and the relative suction zone length, which
is defined as the ratio of the suction zone length to that
at 2-m length.
Effect of Seepage on Sediment Transport Finally, the experimental results also show that the bedload
Candidate: Liu Xiaoxie transport rate in the presence of suction increases rapidly
Report No.: CEE/PhD/2010/215 with time until it reaches a peak, beyond which, the
transport rate decreases. The duration for the bedload
This study presents the experimental results and theoretical transport rate to reach the peak is shorter for flows with
analyses of seepage effects on sediment transport. A total higher sediment transport rates. With an increase in time,
of 529 experiments grouped under four different series of the bedload transport rate reduces, and eventually reaches
tests were conducted in a laboratory flume with a permeable a balance between the incoming and outgoing sediment
sediment boundary to investigate the effect of seepage, the
167
RESEARCH PROJECTS
transport, called the equilibrium condition, i.e., the sediment The objective of this study is to improve the analysis and
transport with no seepage condition. design of flat plate structure slab-column connections by
obtaining representative values of unbalanced moments. It
is achieved by proposing reduced slab stiffness model based
The Role of Microbial Aggregation in Aerobic on the modified effective moment of inertia method. The
Granulation use of the effective moment of inertia method for cracked
Candidate: Luo Yiqun flexural members ensures simplicity and, with appropriate
Report No.: CEE/PhD/2010/216 use of parameters, can lead to accurate prediction of
reduced stiffness of cracked flexural members. Unlike
In the present study, dispersed aerobic granules bacteria currently available models that are only applicable for flat
were subjected to selective hydraulic pressure by settling plate structures with regular columns layouts, the proposed
time, resulting in the loose structured aggregates formed model is also applicable for irregular columns layouts. The
by enrichment process, in which the coaggregating bacteria proposed model has been verified using the experimental
with mild auto aggregating ability were observed as the data of unbalanced moment-lateral drift relationships from
consequence. In comparison, strong auto aggregating both square column-slab connections from past research as
bacteria or co aggregating bacteria only occur in aerobic well as rectangular column-slab connections tested in this
granules sheltered by compact structure. The studies on experimental program. The accuracy of the proposed model
composition and diversity of aggregating microbes in in modeling reduced slab stiffness has been shown much
activated sludge, acetate-fed granule and phenol-degrading better compared to applying uniform reduced slab stiffness,
granule showed that higher proportion of microbial which is commonly used in the available models such as
community members in aerobic granules was involved Effective Beam Width and Equivalent Frame Model. The
in cell-cell aggregating interactions than that in activated proposed model of reduced slab stiffness should be useful
sludge. The fraction of co aggregating microbes can be for design engineers dealing with flat plate slab-column
increased under the chemical and toxic shock. Granule connections to obtain accurate slab deflection, inelastic
isolate S35 possessing both aggregating and autoaggregating lateral drift, and design value of transferred unbalanced
ability was demonstrated to accelerate the formation of moment.
aerobic granule in 5 days. The feasibility of bioaugmentation
of aggregating mono culture to accelerate the granulation This study also investigates the behavior of flat plate
process is supported. Bioaugmentation with granule isolate structure with irregular columns layout, which is currently
S 15 was shown to significantly improve recovery extent of still unclear due to very limited experimental data. The
disintegrated granular sludge. Reaggregated aerobic granules experimental program consists of five corner column-slab
appeared in Rl on day 5 and quickly grew to replace the connections and four slab-column connections with 135-
loose disintegrated granular sludge as a dominant form in degree slabs with rectangular columns, which are often
the biomass. The results suggest the microbial aggregation found in modern flat plate structures. The specimens were
might be an integral part for aerobic granulation, and the tested to investigate their behaviors in term of strength,
presence of aggregating bacteria acted as a trigger to the drift capacity, stiffness, ductility, and the effect of shear
aerobic granulation. reinforcement. The experimental results show the effects
of bidirectional lateral load, gravity load magnitude, and
the use of stud shear reinforcement (SSR).
Behavior of Comer Column-Slab Connections
in Irregular Flat Plate Floors under Gravity Estimation of foreign exchange exposure in
and Bidirectional Lateral Loading Public Private Partnership infrastructure
Candidate: Soerya Widjaja
Report No.: CEE/PhD/2010/217 projects
Candidate: Matthias Ehrlich
The structural behaviors of rectangular column-slab Report No.: CEE/PhD/2010/218
connections in modem flat plate structures are still unclear

due to currently limited experimental data. One of the Economic foreign exchange (FX) exposure is an important
unclear aspects is the design value of transferred unbalanced risk factor which affects Public-private partnership (PPP)
moments, which are critical contributor to the connection projects in developing countries. The risk exists because
shear stresses. More accurate values of unbalanced moments PPP projects typically sell their outputs domestically and
can be obtained by modeling reduced slab stiffness due to generate revenues in local currency, while their financing
slab cracking more accurately. However, currently available costs and operating and maintenance costs are often
models such as Effective Beam Width, Equivalent Frame denominated in hard currencies. Traditionally, FX risk is
Model, and simplified frame analysis, which are only tested through the use of risk factors on revenue and costs
applicable for flat plate structures with regular columns or by adopting conservative assumptions in the cash flow.
layouts, do not represent the actual behavior of the cracked While this method provides a range of a risk value based on
slab. scenarios, it does not give the potential FX risk exposure.
168
RESEARCH PROJECTS
What constitutes minimum and maximum risk values is Stress Concentration Factor and Hot Spot
often defined on the basis of subjective judgements. Stress Studies of Partially Overlapped
Circular Hollow Section K-Joint
This research contributes to the solution of this problem
Candidate: Sopha Thong
with a methodology to quantify annual economic FX
exposure in project companies financed under project finance
modality. The application of the developed FX index to
In the current research, carefully planned experimental
describe the project feasibility on economic FX exposure
studies were firstly carried out to investigate the SCF and the
is superior as it is an extra tool which is linked to the
HSS distributions along the joint intersection of three full-
financial models without the ambiguities to incorporate
scale partially overlapped CHS K-joints. The experimental
risk factors in the cash flow. It is a unique mathematical
results show that depends on the geometrical parameters of
process for dimensioning currency risk on a various set of
the partially overlapped K-joints, the maximum SCF could
cash flow positions.
locate on either the brace side or the chord side of the joint.
The experimental results show that Efthymiou formulae are
A first-order second-moment reliability method based on
conservative only when the partially overlapped K-joints
the Hashofer-Lind reliability index beta was undertaken
are subjected to IPB loading, but not for the case of AX
to reflect the uncertainties of market risks with impact
loading. In addition, it is observed that the S-N curves are
on the cash flow of the PPP project. The FX index was
found to be on the conservative side of the test results. A
modelled via an expanding dispersion ellipsoid in the
comparison between tests results with FE analyses shows
original space of random variables. The input variables
that reliable SCF and HSS values could be obtained.
in the proposed foreign exchange exposure (FEE) model
include inflation rates, interest rates and foreign exchange
After the experimental study, FE models were created to
rates. The variables form the ellipsoid of an n-dimensional
simulate the test specimens. A total of 3500 FE models
shape. It reflects not only the effect of the mean values but
with wide range of geometrical parameters of partially
also the covariances of the random variables influencing a
overlapped K-joints were created. A set of parametric
defined investability domain. The computation of the FX
equations was subsequently proposed for predicting the
index involved eigenvalues and eigenvectors, rotation of
SCF of partially overlapped CHS K-joints. However, as
the reference frame, and transformed space for the random
part of an attempt to gain more an accurate prediction, a
variables.
new method is created for estimating the SCF and HSS
values of partially overlapped CHS K-joint. The assessment
Additionally, a country reliability risk (CRR) index was
confirms the efficiency and reliability of the new method for
designed to evaluate risk mitigation instruments (RMIs).
predicting the SCF and HSS of partially overlapped CHS
FX risk exposure is often mitigated by RMIs. The value of
K-joints under basic AX, IPB and combined loadings.
RMIs depends on the affordability and the willingness of
the government to compensate unforeseen FX fluctuation
Finally, a comparison study between the gapped and the
in the project. Factors influencing country reliability can be
partially overlapped CHS K-joints has been made. It is
identified in the ability to repay debt obligations, liquidity
observed from the comparison results that the partially
difficulties and political difficulties.
overlapped CHS K-joints are mainly in favour when they
are working under the AX load case, while the gapped
Both methodologies the FX index and the CRR index
CHS K-joints are in favour on working under the IPB
represent strategic components in the set of quantitative
loading.
tools. The models can be used as a monitoring tool for
performing FX exposure analysis. It is a forward looking
approach which indicates how prepared the project is in
absorbing economic FX exposure. The models can take Development of Neural Networks in Civil
care of the different institutional arrangements and payment Engineering Applications
structures because they are directly linked to the financial Candidate: Zhang Yun
models. The FX index reflects the annual life cycle costs Report No.: CEE/PhD/2010/220
and revenue structures during the whole concession period.
The outcome illustrates to project sponsors and lenders the The architecture of neural networks (NNs) has a
critical variables that they need to control. The CRR index significant impact on a network’s generalization ability.
provides the default probabilities on RMIs. Both models Ensemble neural networks (ENNs) are commonly used
can be applied to infrastructure projects such as power, networks in many engineering applications due to their
water supply, and transportation. better generalization properties. An ENN usually includes
several back-propagation networks in its structure, where
the back-propagation network is a single feed-forward
network trained with the back-propagation learning rule.
In this thesis, the Akaike information criterion (AIC) and
the entropy were used as the automating design tools for
169
RESEARCH PROJECTS
balancing the generalization against the parameters and the resulting mixing behavior as well as the potential
finding the best combining weights of the ENNs. Two influence of the close proximity of the seabed on the brine
ENNs, namely, the AIC based ENN and the entropy based plume, i.e., the Coanda effect. Assuming flat seabed and a
ENN were developed first. Since the AIC and entropy have combination of mean and oscillatory tidal current, I further
their own merits for solving different problems, a new developed a mathematical model to simulate the long-term
AIC-entropy based ENN was developed. Two analytical salinity build-up on the far field around the outfall due to
functions – the peak function and Friedman function were continuous brine discharge.
used first to assess the accuracy of the proposed ensemble
approaches. The verified approaches were then applied to
the civil engineering applications. Numerical Simulation of Heterogeneous
Material Failure by using the Smoothed
Particle Hydrodynamics Method
Strut-and-Tie Modelling on Deep Beams Candidate: Wang Xuejun
Candidate: Zhang Ning Report No.: CEE/PhD/2010/223
In the thesis, a micromechanical approach based on the
A modified strut-and-tie model for deep beams was Smoothed Particle Hydrodynamics method was developed
developed. Several significant improvements were made and to simulate the heterogeneous material failure by capturing
gave rise to better prediction performance of deep beam the detailed occurring sequence of the microscopic cracks
shear strength, evaluated by 233 test results. No empirical as well as the macro mechanical response. The program
stress limit was required and the concrete softening effect employs an elasto-plastic damage model and utilizes
was embedded in the model. The proposed model was further the statistical approach to account for the heterogeneous
extended to calculate continuous deep beams with rigid and strength distribution in the material microstructure.
elastic supports, which was validated by 54 beams from Besides, a polymineral method that is suitable to model
literature as well as the author’s experimental programme the microstructure of multiphase material with different
on continuous deep beam subjected to differential support components was proposed and implemented in the program.
settlement. The effects of support settlement were studied A series of 2-D and 3-D simulations on rock-like material
and discussed through crack patterns, failure loads, steel failures were performed. The effects of strain rate as well
strains and load-deflection responses. The author further as the material heterogeneity on the specimen fracture
developed the model into a generalised form, including process and its dynamic strength were investigated.
asymmetrical loading conditions and the capability of Comparisons with the experimental results demonstrate
predicting failure modes. Eight beams were tested and good agreements qualitatively. Results also reveal that the
conclusions were drawn on the effects of unsymmetrical strain rate dependency of the dynamic strength might be
loadings on the beam behaviour. The strut-and-tie model ascribed to the apparent confining pressure during the rapid
was also developed to account for size effect in deep beams, loading as well as the material heterogeneity.
followed by a test programme of eleven geometrically-
similar specimens with varied sizes. The proposed model
outperformed several other methods and is a promising
tool for engineers.
Desalination Discharges in Shallow Coastal

Waters
Candidate: Shao Dongdong
From an environmental viewpoint, the outfall of a

desalination plant needs to be properly designed to ensure
the discharged brine plume to be mixed rapidly with the
ambient coastal waters. In this study, I investigated the
behavior of submerged round brine discharges injecting
at 0°, 30° and 45° to stationary receiving water in the
laboratory. Advanced laser diagnostic method of combined
Particle Image Velocimetry (PIV) and Planar Laser Induced
Fluorescence (PLIF) was adopted to measure the velocity
and concentration fields, respectively. The experimental
results obtained provide regulatory references on the gross
geometrical characteristics of the smaller-angle discharges,
170
PUBLICATIONS
PUBLICATIONS
Publications of academic staff in journals and conference proceedings during the period from 2009 to 2010.
Authors who are not members of the School are indicated by *.
“Protection of Structures against Hazards IV”, 2009 - Ed. Han Linhai* and Lok, T.S., CI-Premier, Beijing, P.R. China, 440
pages. (ISBN 978-981-08-3244-5).
“Shock and Impact Loads on Structures VIII”, 2009 - Ed. Wu, C.Q.* and Lok, T.S., CI-Premier, Adelaide, Australia, 791 pages.
(ISBN 978-981-08-3245-2).
Annamdas, V.G.M. and Soh, C.K., 2010. “Application of electromechanical impedance technique for engineering structures:
Review and future issues”. Journal of Intelligent Material Systems and Structures, Vol. 21, No. 1, pp. 41-59.
Bai, H.W., Shao, J.H., Zhang, X.W. and Sun, D.D., 2010. “Effect of TiO2 photocatalytic oxidation for control of membrane
fouling by humic acid in water”. Chinese Journal of Environmental Engineering, Vol. 4, No. 1, pp. 28.
Bao, H.R. and Zhao, Z.Y., 2009. “Indeterminacy of the vertex-vertax contact in 2D discontinuous deformation analysis.”
Proceedings of the 9th International Conference on Analysis of Discontinuous Deformations (ICADD-9), Singapore, pp. 99-
107.
Bao, H.R. and Zhao, Z.Y., 2009. “Modelling crack propagation with nodal-based discontinuous deformation analysis.” Proceedings
of the 9th International Conference on Analysis of Discontinuous Deformations (ICADD-9), Singapore, pp. 161-167.
Bao, H.R. and Zhao, Z.Y., 2010. “An alternative scheme for the corner-corner contact in the two dimensional discontinuous
deformation analysis.” Journal of Advances in Engineering Software, Vol. 41, No. 2, pp. 206-212.
Bao, X.L. and Li, B., 2010. “Residual strength of blast damaged reinforced concrete columns”. International Journal of
Impact Engineering, Vol. 37, pp. 295-308.
Beppu, M.*, Ohno, T.*, Ohkubo, K.*, Li, B. and Satoh, K.*, 2010. “Explosive-resistant performance of fiber sheet reinforced
concrete plates under contact explosion”. International Journal of Protective Structures, Vol. 1, No. 2, pp. 257-270.
Bo, M.W.,* Chu, J., Arulrajah, A.* and Fabius, M.*, 2009. “A case study on predicting primary consolidation settlement
applying small, large strain and stress path methods.” Proceedings of the 17th International Conference on Soil Mechanics and
Geotechnical Engineering, Alexandria, Egypt, 5-9 October.
Brownjohn, J.M.W.* and Pan, T.-C., 2010. “Vibration serviceability of tall buildings due to wind loads: prediction, measurement
and evaluation.” Proceedings of the Structures Congress 2010, 12-15 May 2010, Orlando, Florida, USA.
Chan, C.L. and Low, B.K., 2009. “Reliability analysis of laterally loaded piles involving nonlinear soil and pile behavior.”
Journal of Geotechnical and Geoenvironmental Engineering, ASCE, Vol. 135, No. 3, pp. 431-443.
Cheng, N.S., 2009. “Comparison of formulas for drag coefficient and settling velocity of spherical particles.” Powder
Cheng, N.S., 2009. “Application of incomplete similarity theory for predicting bed-material load discharge.” Journal of
Hydraulic Engineering, ASCE. doi:10.1061/(ASCE)HY.1943-7900.0000375.
Cheng, N.S. and Nguyen, H.T., 2010. “Hydraulic radius for evaluating resistance induced by simulated emergent vegetation
in open channel flows.” Journal of Hydraulic Engineering, ASCE. doi:10.1061/(ASCE)HY.1943-7900.0000377.
Cheng, N.S., Nguyen, H.T., Zhao, K. and Tang, X.*, 2010. “Evaluation of flow resistance in smooth rectangular open-channels.”
Journal of Hydraulic Engineering, ASCE, doi: 10.1061/(ASCE)HY.1943-7900.0000322.
Chiang C. Mei*, Mikhael Krotov*, Zhenhua Huang and Aode Huhe*, 2010. “Short and long waves over a muddy seabed”.
Journal of Fluid Mechanics, Vol. 643, pp. 33-58.
Chiew, S.P. and Beh, C.T.*, 2010. “Use of alternative steel in building steelwork design to BS5950”. International Journal
of Advances in Structural Engineering, Vol. 13, No. 3, pp. 431-439.
171
PUBLICATIONS
Chiew, S.P., Lee C.K., Lie, S.T. and Nguyen, T.B.N.*, 2009. “Fatigue study of partially overlapped circular hollow section
K-Joints”. Proceedings of the 6th International Symposium on Advances in Steel Structures, ICASS-09, Hong Kong, 16-18
December, Vol. I, pp. 602-619.
Chiew, S.P., Lee, C.K. and Lie, S.T., 2009. “Research on tubular joints at Nanyang Technological University”. The IES Journal
Part A – Civil and Structural Engineering, Vol. 2, No. 1, pp. 68-84.
Chiew, S.P., Lee, C.K., Lie, S.T. and Nguyen, T.B.N.*, 2009. “Fatigue study of partially overlapped CHS K-Joints. Part II:
Experimental study and validation of numerical models”. International Journal of Engineering Fracture Mechanics, Vol. 76,
No. 15, pp. 2408-2428.
Chiew, S.P. and Wada, Y.*, 2009. “Corrosion control for building structural steelworks”. Proceedings of the International
Symposium on Advances in corrosion Protection to Steel Members in Building Construction, Hong Kong, 2 November, pp.
103-137.
Chiew, S.P. and Yu, Yi, 2009. “Debonding Behavior of CFRP Strengthened Steel Beams under Static and Cyclic Loads”.
Proceedings of the 6th International Symposium on Advances in Steel Structures, ICASS-09, Hong Kong, 16-18 December,
Vol. I, pp. 285-292.
Chin, K.B., Leong, E.C. and Rahardjo, H., 2009. “Cyclic behaviour of unsaturated silt in suction-controlled simple shear”.
Proceedings of the 4th Asia-Pacific Conference on Unsaturated Soils, Newcastle, Australia, 23-25 November, pp. 65-70.
Chin, K.B., Leong, E.C. and Rahardjo, H., 2010. “A simplified method to estimate soil-water characteristic curve”. Canadian
Geotechnical Journal (accepted for publication).
Chou, S., Shi, L.*, Wang, R., Tang, C.Y., Qiu, C. and Fane, A.G., 2010. “Characteristics and potential applications of a novel
forward osmosis hollow fiber membrane”. Desalination, Vol. 261, pp. 365-372.
Chu, J., Ivanov, V., Lee, M.F., Oh, X.M. and He, J., 2009. “Soil and waste treatment using biocement”. Proceedings of the
International Symposium on Ground Improvement Technologies and Case Histories, 9-11 December, Singapore, Eds, Leung,
C.F., Chu, J., and Shen, R.F., Research Publishing, pp. 160-166.
Chu, J., Varaksin, S.*, Klotz, U.* and Mengé, P.*, 2009. “Construction processes.” State-of-the-Art-Lecture, Proceedings of
the 17th International Conference on Soil Mechanics and Geotechnical Engineering, Alexandria, Egypt, 5-9 October, Vol. 4,
pp. 3006-3135.
Chu, J., Yan, S.W.* and Guo, W., 2009. “Innovative dike construction methods.” Keynote Lecture, International Symposium
on Geotechnical Engineering, Ground Improvement & Geosynthetics for Sustainable Mitigation and Adaptation to Climate
Change including Global Warming, 3-4 December, Bangkok, Thailand.
Chua, H.C., Goh, A.T.C. and Zhao, Z.Y., 2009. “Distinct element analysis of stage constructed underground cavern in the
vicinity of a fault.” Proceedings of the 9th International Conference on Analysis of Discontinuous Deformation (ICADD-9),
Singapore, pp. 429-435.
Chua, L.H.C., Leong, M.C.M., Lo, E.Y.M., Reinhard, M., Robertson, A.P.*, Lim, T.T., Shuy, E.B. and Tan, S.K., 2009. “Controlled
field studies on artificial recharge by surface infiltration in a sandfill”. Water Science and Technology, IWA publishing, Vol.
60, No. 5, pp. 1283-1293.
Chua, H.C.L., Lo, Y.M.E., Shuy, E.B. and Tan, B.K.S., 2009. “Nutrients and suspended solids in storm runoff from catchments
with various proportions of rural and urban land use in Kranji catchment, Singapore”. Journal of Environmental Management,
Vol. 90, pp. 3635-3642.
Chua, L.H.C., Lo, E.Y.M., Lim, T.T., Robertson, A.P.*, Shuy, E.B. and Tan, S.K., 2009. “Geochemical changes during recharge
with tertiary-treated wastewater at a coastal sandfill”. Water Science and Technology, IWA Publishing, Vol. 60, No. 5, pp.
1273-1281.
Chua, L.H.C. and Wong, T.S.W., 2010. “Improving event-based rainfall-runoff modelling using a combined artificial neural
network-kinematic wave approach”. Journal of Hydrology, Vol. 390, pp. 92-107.
Chua, L.H.C., Lo, E.Y.M., Shuy, E.B., Robertson, A.P.*, Lim, T.T. and Tan, S.K., 2010. “DOC and UVA attenuation with
soil aquifer treatment in the saturated zone of an artificial coastal sandfill”. Water Science and Technology, Vol. 62, No. 3,
pp. 491-500.
172
PUBLICATIONS
Chua, L.H.C., Wong, T.S.W. and Wang, X.H., 2010. “Determination of the constant loss rate and other physical parameters
in event-based rainfall-runoff modelling using linear artificial neural networks”. Applied Soft Computing, Vol. 11, No. 1, pp.
373-381.
Dao, M.H., Tkalich, P.*, Chan, E.S. and Megawati, K., 2009. “Tsunami propagation scenarios in the South China Sea.” Journal
of Asian Earth Sciences, Vol. 36, pp. 67-73.
Ding, H.-B., Tan, G.-Y. Amy, and Wang, J.-Y., 2010. “Caproate formation in mixed-culture fermentative hydrogen production”.
Bioresource Technology (Doi: 10.1016/j.biotech. 2010.07.056).
Feng, C.S.*, Wang, R., Wu, Y.* and Li, G.*, 2010. “Preliminary analysis of a linear pore pattern formed on poly (vinylidene
fluoride-co-hexafluoro propylene) porous membrane surfaces”. Journal of Membrane Science, Vol. 352, pp. 255-261.
Feng, C.S.*, Wang, R., Zhang, H.Y.* and Shi, L.*, 2010. “Diverse morphologies of PVDF hollow fibre membranes and their
performance analysis as gas/liquid contactors”. Journal of Applied Polymer Science, in press.
Foreman, J.*, Gallien, J.*, Alspaugh, J.*, Lopez, F.*, Bhatnagar, R.*, Teo, C.C. and Dubois, C.* 2010. “Implementing supply-
routing optimization in a make-to-order manufacturing network.” Manufacturing and Service Operations Management, Vol.
12, pp. 547-568.
Fujikake, K.*, Li, B. and Soeun, S.*, 2009. “Impact response of reinforced concrete beam and its analytical evaluation”. ASCE
Journal of Structural Engineering, August, Vol. 135, No. 8, pp. 938-950.
Gao, Y.Y.*, Yu, D.Y.*, Tan, S.K., Wang, X.K. and Hao, Z.Y., 2010. “Experimental study on the near wake behind two side-
by-side cylinders of unequal diameters”. Fluid Dynamic Research, 42 055509.
Gao, Y.Y., Stephane Etienne,* Yu, D.Y.* and Tan, S.K., 2010. “Flow characteristics behind two unequal circular cylinders
in tandem arrangement”. ISOPE-2010 ¾ The 20th International Offshore and Polar Engineering Conference, Beijing, China,
20-26 June 2010.
Gao, Y.-Y.*, Yu, D.-Y.*, Tan, S.K., Wang, X.K. and Hao, Z., 2010. “Flow behaviour behind two side-by-side circular cylinders
with unequal diameters”. Proceedings of the 29th International Conference on Ocean, Offshore and Arctic Engineering
(OMAE2010), 6-11 June, Shanghai, China, OMAE2010-20217.
Gensheimer, R.J.*, Wang R.Q.*, Adams, E.E.*, Daichin*, Shao, D., Zhao, B., Huang, Z. and Law, A.W.K., 2010. “Dynamics
of particle clouds with application to open water sediment disposal.” Proceedings of the 6th International Symposium on
Environmental Hydraulics, 23-25 June, Athens, Greece.
Giannis, A.*, Pentari, D.*, Wang, J.Y. and Gidarakos, E.*, 2010. “Application of sequential extraction analysis to electrokinetic
remediation of cadmium, nickel and zinc from contaminated soils”. Journal of Hazardous Materials, Vol. 184, pp. 547-554.
Goh, A.T.C. and Hefney, A.M.*, 2010. “Reliability assessment of EPB tunnel-related settlement.” International Journal
Geomechanics and Engineering, Vol. 2, No. 1, pp. 57-69.
Goh, S.G., Rahardjo, H. and Leong, E.C., 2009. “Evaluation of shear strength equations for unsaturated soil”. Proceedings
of the 4th Asia-Pacific Conference on Unsaturated Soils, Newcastle, Australia, 23-25 November, pp. 753-758.
Goh, S.G., Rahardjo, H. and Leong, E.C., 2010. “Shear strength equations for unsaturated soil under drying and wetting”.
ASCE Journal of Geotechnical and Geoenvironmental Engineering, April, Vol. 136, No. 4, pp. 594-606.
Gu, J. and Zhao, Z.Y., 2009. “Considerations of the discontinuous deformation analysis on wave propagation problems.”
International Journal of Numerical and Analytical Methods in Geomechanics, Vol. 33, No. 12, pp. 1449-1465.
Guo, W., Chu, J. and Yan, S.W.*, 2009. “Classification of geotubes and related analysis methods”. Proceedings of the
International Symposium on Ground Improvement Technologies and Case Histories, Eds, Leung, C.F., Chu, J. and Shen, R.F.,
Research Publishing, pp. 263-274.
Hao, Z., Zhou, T., Wang, X.K. and Tan, S.K., 2010. “Experimental studies of vortex structures in the wake of a cylinder with
helical strakes”. Proceedings of the 29th International Conference on Ocean, Offshore and Arctic Engineering (OMAE2010),
Hay, C.T., Khor, S.L., Sun, D.D. and Leckie, J.O.*, 2009. “Influence of a prolonged solid retention time environment on
nitrification/denitrification and sludge production in a submerged membrane bioreactor”. Desalination, Vol. 245, pp. 28-43.
(IF: 1.155).
173
PUBLICATIONS
He, J.W. and Low, Y.M., 2010. “Probabilistic assessment of the clashing between flexible marine risers”. Proceedings of the
International Conference on Ocean, Offshore and Arctic Engineering, Shanghai, China.
He, L.*, Huang, G.H.* and Qin, X.S., 2010. “An integrated distributed-hydrologic and watershed-management model: A case
study in the Heshui River Watershed of Southern China.” Proceedings of the 4th International Yellow River Forum (IYRF) on
Ecological Civilization and River Ethics, Zhengzhou, China, October 20-23, 2009, Vol. 1, pp. 197-205.
Hota, G.*, Sundarrajan, S.*, Ramakrishna, S.* and Ng, W.J., 2009. “One step fabrication of MgO solid and hollow submicrometer
fibers via electrospinning method”. J. Am. Ceram. Soc., Vol. 92, No. 10, pp. 2429-2433.
Huang, G. and Law, A.W.K., 2009. “Taylor dispersion under random waves.” Proceedings of the 5th International Conference
on. Asian and Pacific Coasts (APAC2009), 13-16 October, Singapore.
Huang, G., Law, A.W.K. and Huang, Z., 2010. “Experimental study on wave-induced drift of small floating.” Proceedings of
the 9th International Conference on Hydro-Science and Engineering, 2-5 August, Chennai, India.
Huang, G.H.*, Sun, W.*, Nie, X.H.*, Qin, X.S. and Zhang, X.D.*, 2010. “Development of a decision support system for
rural eco-environmental management in Yongxin County, Jiangxi Province, China.” Environmental Modelling and Software,
Vol. 25, No. 1, pp. 25-42.
Huang, Z., Wu, T.-R., Tan, S.K., Megawati, K., Shaw, F., Liu, X. and Pan, T.-C., 2009. “Tsunami hazard from the subduction
megathrust of the South China Sea: Part II. Hydrodynamic modelling and possible impact on Singapore.” Journal of Asian
Earth Sciences, Vol. 36, pp. 93-97.
Huang, Z.H. and Aode, H., 2009. “A laboratory study of rheological properties of mudflows in Hangzhou Bay, China”.
International Journal of Sediment Research, Vol. 24, pp. 409-423.
Huang, Z.H., Liu, C.R., Adi, K., Tan, S.K. and Nah, E., 2009. “Responses of a floating rectangular caisson to regular waves:
A comparison of measurements with time-domain and frequency-domain simulations”. Proceedings of the 5th International
Conference on Asian and Pacific Coasts (Paper No. 200-206). Singapore: World Scientific Publishing.
Huang, Z.H., 2010. “A note on tsunami hazard mitigation by mangrove forests”. Proceedings of the 9th International Conference
on Hydroinformatics (Paper No. HICA00455-00733), China: Chemical Industry Press.
Huang, Z.H. and Yuan, Z.D., 2010. “Transmission of solitary waves through slotted barriers: A laboratory study with analysis
by a long wave approximation”. Journal of Hydro-Environment Research, Vol. 3, No. 4, pp. 179-185.
Huang, Z.H. and Zhang, W.B., 2010. “A laboratory study of dynamic responses of a moored rectangular floating breakwater
to regular waves” (7 pages). The International Symposium on Hydraulic Physical Modeling and Field Investigation. 13-15
September 2010, Nanjing, China.
Indrawan, I.G.B. and Rahardjo, H., 2010. “Water infiltration through capillary barrier models.” Proceedings of the Symposium
on “Protecting Life from Geo-Disaster and Environmental Hazards”, Department of Geological Engineering, Gadjah Mada
University and AUN/SEED-Net JICA, Bali, Indonesia, 25-26 February, pp. 439-446 (D14-(1-8).
Jia, Y., Wang, R. and Fane, A.G., 2009. “Hybrid PAC-submerged membrane system for trace organics removal I. Adsorption
kinetics study of PAC in a bubbled solution”. Chemical Engineering Journal, Vol. 155, pp. 155-160.
Jia, Y., Wang, R. and Fane, A.G., 2009. “Hybrid PAC-submerged membrane system for trace organics removal II: System
simulation and application study”. Chemical Engineering Journal, Vol. 149, pp. 42-49.
Jiang, B. and Liu, Y., 2010. “Energy uncoupling inhibits aerobic granulation”. Applied Microbiology and Biotechnology, Vol.
85, No. 3 pp. 589-595.
Jiang, H.-L., Maszenan, A.M., Zhao, Z.-W. and Tay, J.-H., 2010. “Properties of phenol-removal aerobic granules during normal
operation and shock loading”. Journal of Industrial Microbiology & Biotechnology, Vol. 37, pp. 253-262.
Jiang, X., Zhou, Y.* and Ng, W.J. 2010. “Acidogenic removal of monochlorophenols”. Proceedings of the 12th World Congress
on Anaerobic Digestion, 31 October – 4 November 2010, Guadalajara, Mexico.
Jinadasa, K.B.S.N.*, Sasikala, S.*, Tanaka, N.*, Mowjood, M.I. M.* and Ng, W.J., 2009. “Effect of pulsing application on
performance of tropical constructed wetland treating domestic waste”. International Symposium on Southeast Asian Water
Environment, Vol. 7, pp. 435-442.
174
PUBLICATIONS
Kim, W.*, Lee, S.*, Shin, S.G.*, Lee, C., Hwang, K.* and Hwang, S.*, 2010. “Monitoring methanogenic community changes
in duplicate anaerobic batch digesters treating swine wastewater”. Water Research, Vol. 44, pp. 4900-4907.
Krauthammer, T.*, Langseth, M.*, Ohno, T.*, Thoma, K.*, Pan, T.-C., and Lim, C.H., 2010. “Design and analysis of protective
structures – advances in protective technology”. Proceedings of the 3rd International Conference on Design and Analysis of
Protective Structures 2010 (DAPS-2010), Defence Science & Technology Agency, Singapore, 10-12 May 2010, Singapore.
Krisdani, H., Rahardjo, H. and Leong, E.C., 2010. “Application of geosynthetic material as a coarse-grained layer in capillary
barriers”. Special Issue on Unsaturated Geosynthetics, Geosynthetics International Journal, Vol. 17, No. 5, pp. 323-331.
Krisdani, H., Rahardjo, H. and Leong, E.C., 2010. Response to Discussion by Mbonimpa, M., Aubertin, M. and Bussiere, B.
on “Effects of different drying rates on shrinkage characteristics of a residual soil and soil mixtures”. Journal of Engineering
Geology, Vol. 107, No. 3-4, pp. 172-173 (2009). Published in Journal of Engineering Geology, Vol. 110, No. 1-2, pp. 30-
31.
Kulkarni, S.A. and Li, B., 2009. “Investigations on seismic behaviour of hybrid-steel concrete connections”. Journal of
Precast/Prestressed Concrete Institute (PCI), Winter, Vol. 54, pp. 67-87.
Kulkarni, S.A. and Li, B., 2009. “Seismic behaviour of reinforced concrete interior wide beam-column joints”. Journal of
Earthquake Engineering, January, Vol. 13, Issue 1, pp. 80-99.
Kumara, C.K.*, Ng, W.J., Bandara, A.* and Weerasooriya, R.*, 2010. “Nanogibbsite: Synthesis and characterization”. Journal
of Colloid and Interface Science, Vol. 352, pp. 252-258.
Kuniawan, A., Huang, Z.H., Li, J., Liu, C.R.*, Wang, X.K., Hao, Z.Y., Tan, S.K. and Nah, E.*, 2009. “A numerical analysis
of the response and air gap demand for semi-submersibles”. Proceedings of the 28th International Conference on Ocean,
Offshore and Arctic Engineering, OMAE2009, 31 May - 5 June 2009, Honolulu, Hawaii, USA, OMAE2009-79163.
Law, A.W.K., Au, S.* and Song, J., 2010. “Stochastic diffusion by progressive waves in turbulence.” Proceedings of the 9th
International Conference on Hydrodynamics, 11-15 October, Shanghai, China.
Lay, W.C.L., Chong, T.H., Tang, C.Y., Fane, A.G., Zhang, J. and Liu, Y., 2010. “Fouling propensity of forward osmosis:
investigation of the slower flux decline phenomenon.” Water Science and Technology, Vol. 61, No. 4, pp. 927-936.
Lay, W.C.L., Liu, Y. and Fane, A.G., 2010. “Impacts of salinity on the performance of high retention membrane bioreactors
for water reclamation: A review”. Water Research, Vol. 44, No. 1, pp. 21-40.
Lee, C., Kim, J.*, Chinalia, F.A.*, Shin, S.G.* and Hwang, S.*, 2009. “Unusual bacterial populations observed in a full-scale
municipal sludge digester affected by intermittent seawater inputs”. Journal of Industrial Microbiology and Biotechnology,
Vol. 36, pp. 769-773.
Lee, C., Kim, J.*, Hwang, K.* and Hwang, S.*, 2009. “Fermentation and growth kinetic study of Aeromonas caviae under
anaerobic conditions”. Applied Microbiology and Biotechnology, Vol. 83, pp. 767-773.
Lee, C., Kim, J.*, Hwang, K.*, O’Flaherty, V.* and Hwang, S.*, 2009. “Quantitative analysis of methanogenic community
dynamics in three anaerobic batch digesters treating different wastewaters”. Water Research, Vol. 43, pp. 157-165.
Lee, C., Kim, J.*, Shin, S.G.*, O’Flaherty, V.* and Hwang, S.*, 2010. “Quantitative and qualitative transitions of methanogen
community structure during the batch anaerobic digestion of cheese-processing wastewater”. Applied Microbiology and
Biotechnology, Vol. 87, pp. 1963-1973.
Lee, C.K., Chiew, S.P., Lie, S.T. and Nguyen, T.B.N.*, 2009. “Fatigue study of partially overlapped circular hallow section-
joints, Part I: Geometrical models and mesh generation.” Engineering Fracture Mechanics, Vol. 76, No. 16, pp. 2445-2463.
Lee, C.K., Chiew, S.P., Lie, S.T. and Nguyen, T.B.N.*, 2010. “Adaptive mesh generation procedures for thin-walled tubular
structures.” Finite Element in Analysis and Design, Vol. 46, No.1-2, pp. 114-131.
Lee, C.K., Chiew, S.P., Lie, S.T., Sopha, T.* and Nguyen, T.B.N.*, 2009. “Experimental studies on stress concentration factors
for partially overlapped circular hollow section K-Joints”. International Journal of Advanced Steel Construction, Vol. 5, No.
4, pp. 481-499.
Lei, L., Bai, H.W., Liu, Z.Y. and Sun, D.D., 2010. “Hierarchical Ag/TiO2 nanofiber membrane for water purification”. IWA
LET Conference, Phoenix, USA, 3 June 2010.
175
PUBLICATIONS
Leong, E.C., He, L.C. and Rahardjo, H., 2009. Discussion on “Assessment of the use of the vapour equilibrium technique
in controlled-suction tests” by Pintado, X., Lloret, A. and Romero, E., (2009). Canadian Geotechnical Journal, Vol. 46, pp.
411-423. Published in Canadian Geotechnical Journal, Vol. 46, pp. 1482-1484.
Leong, E.C., Rahardjo, H. and He, L.C., 2010. Discussion on “Calibrations of a high-suction tensiometer” by Laurenco, S.D.N.,
Gallipoli, D., Toll, D.G., Augarde, C.E., Evans, F.D. and Medero, G.M. (2008), Geotechnique, Vol. 58, No. 8, pp. 659-668.
Published in Geotechnique, Vol. 60, No. 3, pp. 233-234.
Leung, C.F.*, Chu, J. and Shen, R.F.*, (Editors) 2009. “Ground improvement technologies and case histories”. Research
Publishing, ISBN: 981-08-3124-2.
Li, B. and Chen, Q., “Stiffness of reinforced concrete structural walls with irregular openings”. Journal of Earthquake
Engineering & Structural Dynamics, Vol. 39, Issue 4, pp. 397-417.
Li, B. and Chua, G.H.Y., 2009. “Seismic performance of strengthened reinforced concrete beam-column joints using FRP
composites”. ASCE Journal of Structural Engineering, Vol. 135, No. 10, pp. 1177-1190.
Li, B., Kulkarni, S.A. and Leong, C.L., 2009. “Seismic performance of precast hybrid-steel concrete connections”. Journal
of Earthquake Engineering, June, Vol. 13, Issue 5, pp. 667-689. (First-Tier).
Li, B., Pan, T.-C. and Nair, A., 2009. “A case study of the effect of cladding panels on the response of reinforced concrete
frames subjected to distant blast loadings”. Nuclear Engineering and Design, March, Vol. 239, Issue 3, pp. 455-469.
Li, B., Pan, T.-C. and Tran, C.T.N. 2009. “Effect of axial compression load on seismic behaviour of non-seismically detailed
interior beam-wide column joints”. ACI Structural Journal, September-October, Vol. 106, No. 5, pp. 591-599.
Li, B., Pan, T.-C. and Tran, C.T.N., 2009. “Effects of axial compression load and eccentricity on seismic behaviour of non-
seismically detailed interior beam-wide column joints”. ASCE Journal of Structural Engineering, July, Vol. 135, No. 7, pp.
774-784.
Li, B., Pan, T.-C. and Tran, C.T.N., 2009. “Seismic behavior of nonseismically detailed interior beam-wide column and beam-
wall connections”. Structural Journal, American Concrete Institute, Vol. 106, No. 5, pp. 591-599.
Li, B. and Tran, C.T.N., 2009. “Seismic behaviour of reinforced concrete beam-column joints with vertical distributed
reinforcement”. ACI Structural Journal, November-December, Vol. 106, No. 6, pp. 790-799.
Li, B., Tran, C.T.N. and Pan, T.C., 2009. “Experimental and numerical investigations on the seismic performance of lightly
reinforced concrete joints”. ASCE Journal of Structural Engineering, Vol. 135, No. 9, pp. 1007-1018.
Li, B., Huang, Z.W. and Lim, C.L., 2010. “The verification of non-dimensional energy spectrum based blast design for reinforced
concrete members through actual blast tests”. ASCE Journal of Structural Engineering, June, Vol. 136, No. 6, pp. 627-636.
Li, B. and Kulkarni, S.A., 2010. “Seismic behaviour of reinforced concrete exterior wide beam-column joints”. ASCE Journal
of Structural Engineering, Vol. 136, No. 1, pp. 26-36.
Li, B. and Lim, C.L., 2010. “Tests on seismically damaged reinforced concrete structural walls repaired using fiber-reinforced
polymers”. ASCE Journal of Composites for Construction, No. 10.
Li, B., Pan, T.-C., and Nair, A. “A case study of the local and global structural responses of a tall building in Singapore subjected
to close-in detonations”. The Structural Design of Tall and Special Buildings, John Wiley & Sons, Ltd, UK (to appear).
Li, F.Z. and Low, Y.M., 2010. “Sensitivity study of critical parameters influencing the uncertainty of fatigue damage in steel
catenary risers”. Proceedings of the International Conference on Ocean, Offshore and Arctic Engineering, Shanghai, China.
Li, H.Z.* and Low, B.K., 2010. “Reliability analysis of circular tunnel under hydrostatic stress field.” Computers and
Geotechnics, Vol. 37, Issue 1-2, pp. 50-58.
Li, J., Tan, S.K., Huang, Z.H. and Kurniawan, A., 2009. “Wave amplification and air-gap response under a multi-column
platform”. Coastal Dynamics 2009, 7-11 September 2009, Tokyo, Japan.
Li, J., Huang, Z.H. and Tan, S.K., 2010. “Extreme air-gap response below deck of floating structures”. The International
Journal of Ocean and Climate Systems, Multi-Science Publishing, March, Vol. 1, No. 1, pp. 15-26.
176
PUBLICATIONS
Li, J., Liu, H.X. and Tan, S.K., 2010. “Lagrangian modelling of tidal bores passing through bridge piers”. Proceedings of
the 9th International Conference on Hydrodynamics (ICHD2010) (in press).
Liang, D.W.*, Shayegan, S.S.*, Ng, W.J. and He, J.Z.*, 2010. “Development and characteristics of rapidly formed hydrogen-
producing granules in an acidic anaerobic sequencing batch reactor (AnSBR)”. Biochemical Engineering Journal, Vol. 49,
pp. 119-125.
Lv, L.*, Lu, Y.Q.*, Ng, W.J. and Zhao, X.S.*, 2009. “Bactericidal activity of silver nanoparticles supported on microporous
titanosilicate ETS-10”. Microporous and Mesoporous Materials, Vol. 120, No. 3, pp. 304-309.
Lie, S.T. and Yang, Z.M., 2009. “BS7910: 2005 Failure assessment diagram (FAD) on cracked circular hollow section (CHS)
Welded Joints”. International Journal of Advanced Steel Construction, Vol. 5, No. 4, pp. 385-393.
Lie, S.T. and Yang, Z.M., 2009. “Fracture assessment of damaged square hollow section (SHS) K-joint using BS7910:2005”.
Engineering Fracture Mechanics, Vol. 76, No. 9, pp. 1303-1319.
Lie, S.T. and Yang, Z.M., 2009. “Safety assessment procedure for a cracked square hollow section (SHS) Y-joint”. International
Journal of Advances in Structural Engineering, Vol. 12, No. 3, pp. 359-372.
Lie, S.T. and Yang, Z.M., 2009. “Validation of BS7910: 2005 Failure assessment diagram for cracked square hollow section
T-, Y- and K-joints”. International Journal of Pressure Vessels and Piping, Vol. 86, No. 5, pp. 291-344.
Lie, S.T. and Yang, Z.M., 2010. “Plastic collapse loads of cracked square hollow section (SHS) T-, Y- and K-joints”. Journal
of Offshore Mechanics and Arctic Engineering (OMAE), American Society of Mechanical Engineers (ASME), Vol. 132, No.
3, pp. 1-10.
Lie, S.T. and Zhang, B.F., 2010. “Plastic collapse load investigation for safety assessment of cracked square hollow section
(SHS) T-, Y- and K-joints”. Proceedings of the 29th International Conference on Ocean, Offshore and Arctic Engineering
(OMAE 2010), 6-11 June 2010, Shanghai, China, Paper OMAE2010-20324.
Lie, S.T., Zhang, B.F. and Yang, Z.M., 2010. “Numerical and experimental plastic collapse loads and CTODs of a cracked
square hollow section (SHS) K-joint”. Proceedings of the 13th International Symposium on Tubular Structures, 15-17 December
2010, Hong Kong, China.
Lim, C.L., Li, B. and Pan, T.-C., 2009. “Seismic performances of reinforced concrete frames with wall-like columns”. IES
Journal Part A: Civil and Structural Engineering, Vol. 2, No. 2, May, pp. 126-142.
Lim, Y.Y. and Soh, C.K., 2010. “Estimation of fatigue life using electromechanical impedance technique”. Proceedings of the
SPIE, March, San Diego, USA, Vol. 7647, p. 64722.
Lin, Q.G.*, Huang, G.H.*, Brad, B.*, Nie, X.H.*, Zhang, X.D.* and Qin, X.S., 2010. “EMDSS: An optimization-based decision
support system for energy systems management under changing climate conditions - an application to the Toronto-Niagara
Region, Canada”. Expert Systems with Applications, Vol. 37, No. 7, pp. 5040-5051.
Listiarini, K., Chan, W., Sun, D.D. and Leckie, J.O.*, 2009. “Fouling mechanism and resistance analyses of systems containing
sodium alginate, calcium, alum and their combinations in dead-end fouling of nanofiltration membranes”. Journal of Membrane
Science, Vol. 344, No. 1-2, pp. 244-251. (IF:3.247).
Listiarini, K., Sun, D.D. and Leckie, J.O., 2009. “Cake characterization of sodium alginate fouling of nanofiltration membranes
in the presence of calcium and alum”. IWA Water and Industry 2009, New Zealand, 30 November to 1 December 2009.
Liu, C., Huang, Z.H., and Tan, S.K., 2009. “Nonlinear scattering of non-breaking waves by a submerged horizontal plate:
Experiments and simulations”. Ocean Engineering, Vol. 36, pp. 1332-1345.
Liu, C., Qiu, Q. and Huang, Z.H., 2009. “Higher harmonic waves generated by a submerged horizontal thin plate: An experimental
study for breaking and non-breaking waves”. Proceedings of the 5th International Conference on Asian and Pacific Coasts,
Singapore, World Scientific Publishing, pp. 162-169.
Liu, C., Huang, Z.H., Law, A.W.K. and Geng, N., 2010. “A numerical study of wave energy converter in the form of an
oscillating water column based on a mixed Eulerian-Lagrangian formulation”. Proceedings of the 29th International Conference
on Ocean, Offshore and Arctic Engineering (Paper No. OMAE2010-21056). USA: ASME.
177
PUBLICATIONS
Liu, C.L.*, Li, G.Q.*, Sun, J.Y.* and Lok, T.S., 2009. “Design of planter-box as anti-ram barriers to resist vehicle bomb”.
Proceedings of the 4th International Conference on Protection of Structures against Hazards, Beijing, P.R. China, pp. 245-
253.
Liu, C.L.*, Palermo, D.*, Lok, T.S. and Chen X.L.*, 2009. “An analytical cost methodology in protective solution”. Proceedings
of the 8th International Conference on Shock and Impact Loads on Structures, Adelaide, Australia, pp. 379-388.
Liu, H.L.*, Chu, J. and Ren, Z.Y.*, 2009. “New methods for measuring the installation depth of prefabricated vertical drains.”
Geotextiles and Geomembranes, Vol. 29, No. 6, pp. 493-496.
Liu, H.W., Ghidaoui, M.S., Huang, Z.H. and Yuan, Z., 2009. “Numerical investigation of the interaction between solitary
waves and pile breakwaters”. Proceedings of the 5th International Conference on Asian and Pacific Coasts, Singapore, World
Scientific Publishing, pp. 163-169.
Liu, H.X., Li, J. and Tan, S.K., 2010. “Environmental fluid dynamics – jet flow”. Proceedings of the 9th International
Conference on Hydrodynamics (ICHD2010) (in press).
Liu, H.X., Tan, S.K., Li, J. and Wang, X.K., 2010. “Three dimensional simulation of bore flow using SPH”. Proceedings of the
29th International Conference on Offshore Mechanics and Arctic Engineering (OMAE2010), Shanghai, OMAE 2010-21090.
Liu, J.X., Zhao, Z.Y. and Liang, N.G.*, 2010. “Numerical and theoretical investigations of the tensile failure of shrunk
cement-based composites.” Chapter 2 of Computational Mechanics Research Trends, Editors: Hans P. Berger, Nova Science
Publishers, pp. 111-148.
Liu, J.X., Zhao, Z.Y., Deng, S.C.* and Liang, N.G.*, 2009. “A simple method to simulate shrinkage induced cracking in
cement-based composites by lattice-type modeling.” Computational Mechanics, Vol. 43, No. 4, pp. 477-492.
Liu, J.X., Zhao, Z.Y., Zhang, J.* and Liang, N.G.*, 2009. “Numerical investigation of crack growth in concrete subjected to
compression by the generalized beam lattice model.” Computational Mechanics, Vol. 43, No. 2, pp. 277-295.
Liu, Q.S.*, Liu, Y., Show, K.Y.* and Tay, J.H., 2009. “Toxicity effect of phenol on aerobic granules”. Environmental
Liu, S.S., Zhang, X.W., Sun, D.D. and Xu, Z.M., 2009. “Study on membrane fouling caused by activated sludge from a
membrane bioreactor with long solid retention time”. Chinese Journal of Environmental Engineering, Vol. 2, No. 10, pp.
1816-1820.
Liu, Y., 2009. “Is the free energy change of adsorption correctly calculated?” Journal of Chemical and Engineering Data,
Vol. 54, No. 7, pp. 1981-1985.
Liu, Y.J. and Sun, D.D., 2010. “Comparison of membrane fouling in dead-end microfiltration granular sludge suspension and
its supernatant”. Journal of Membrane Science, Vol. 352, No. 1-2, pp. 100-106. (IF:3.247).
Loh, C.-H.*, Mao, C.-H.*, Huang, J.-H.* and Pan, T.-C. “System identification and damage evaluation of degrading hysteresis
of reinforced concrete frames”. Journal of Earthquake Engineering and Structural Dynamics, International Association for
Earthquake Engineering (to appear).
Low, B.K., 2010. “Slope reliability analysis: some insights and guidance for practitioners.” Proceedings of the 17th Southeast
Asian Geotechnical Conference, Taipei, Taiwan, 10-13 May, Vol. 2, pp. 231-234.
Low, Y.M., 2009. “Frequency domain analysis of a tension leg platform with statistical linearization of the tendon restoring
forces”. Marine Structures, Vol. 22, No. 3, pp. 480-503.
Low, Y.M., 2009. “Efficient vector outcrossing analysis of the excursion of a moored vessel”. Probabilistic Engineering
Mechanics, Vol. 24, No. 4, pp. 565-576.
Low, Y.M., 2009. “Fatigue analysis of deepwater risers using a hybrid time/frequency domain method”. Proceedings of the
International Conference on Offshore and Polar Engineering, Osaka, Japan, Vol. 2, pp. 389-395.
Low, Y.M., 2010. “A method for accurate estimation of the fatigue damage induced by bimodal processes”. Probabilistic
Engineering Mechanics, Vol. 25, No. 1, pp. 75-85.
Low, Y.M., 2010. “Influence of the setdown of a tension leg platform on the extreme airgap response”. Applied Ocean
Research, Vol. 32, No. 1, pp. 11-19.
178
PUBLICATIONS
Low, Y.M., 2010. “A practical formulation for estimating the extreme vector excursion of a floating structure”. Ocean
Engineering, Vol. 37, No. 13, pp. 1159-1168.
Low, Y.M. and Grime, A.J.*, 2010. “Extreme response analysis of floating structures using coupled frequency domain analysis”.
Journal of Offshore Mechanics and Arctic Engineering (accepted).
Low, Y.M. and Grime, A.J.*, 2010. “Extreme response analysis of floating structures using coupled frequency domain analysis”.
Proceedings of the International Conference on Ocean, Offshore and Arctic Engineering, Shanghai, China.
Luo, G., Xie, Li., Zhou, Z.H., Zhou, Q. and Wang, J.Y., 2010. “Fermentative hydrogen production from cassava stillage by
mixed anaerobic microflora: Effects of temperature and pH”. Journal of Applied Energy (accepted).
Ma, H.*, Yang, D.* and Tan, S.K., 2010. “Impacts of climate change and human activities on the flow discharge in the Miyun
Reservoir Catchment”. Journal of Hydrology (2010), doi:10.1016/j.jhydrol.2010.06.010.
McKeown, R.M.*, Scully, C.*, Enright, A.M.*, Chinalia, F.A.*, Lee, C., Mahony, T.*, Collins, G.* and O’Flaherty, V.*, 2009.
“Psychrophilic methanogenic community development during long-term cultivation of anaerobic granular biofilms”. The ISME
Journal, Vol. 3, pp. 1231-1242.
Megawati, K. and Pan, T.-C., 2009. “Regional seismic hazard posed by the Mentawai segment of the Sumatran megathrust.”
Bulletin of Seismological Society of America, Vol. 99, No. 2A, pp. 566-584.
Megawati, K. and Pan, T.-C. 2010. “Ground-motion attenuation relationship for Sumatran megathrust earthquakes”. Journal
of Earthquake Engineering and Structural Dynamics, International Association for Earthquake Engineering, Vol. 39, No. 8,
pp. 827-845.
Megawati, K. and Pan, T.C., 2010. “Development and validation of ground-motion attenuation relationship for large-magnitude
subduction earthquakes.” Proceedings of the 9th U.S. National and 10th Canadian Conference on Earthquake Engineering
(9USN/10CCEE), Toronto, Canada, 25-29 July 2010.
Megawati, K., Shaw, F., Sieh, K., Huang, Z., Wu, T.-R., Lin, Y., Tan, S.K. and Pan, T.-C., 2009. “Tsunami hazard from the
subduction megathrust of the South China Sea: Part I. Source characterization and the resulting tsunami.” Journal of Asian
Earth Sciences, Vol. 36, pp. 13-20.
Mun, C.H.*, He, J.Z.* and Ng, W.J., 2010. “Pentachlorophenol dechlorination by an acidogenic sludge”. Water Research,
Vol. 42, pp. 3789-3798.
Ng, C.A., Sun, D.D., Zhang, J., Wu, B. and Fane, A.G., 2010. “Mechanisms of fouling control in membrane bioreactors by
the addition of powdered activated carbon”. Separation Science and Technology, Vol. 45, No. 7, pp. 873-889. (IF:1.139).
Ng, J.W., Zhang, X.W., Zhang, T., Pan, J.-H., Du, A.J.-H. and Sun, D.D., 2009. “Construction of self-organized free-standard
TiO2 nanotube arrays for effective disinfection of drinking water”. Proceedings of the 2nd European Conference on Environmental
Applications of Advanced Oxidation Processes (EAAOP2), Cyprus, 9-11 September 2009.
Ng, J.W., Zhang, X.W., Zhang, T., Pan, J.-H., Du, A.J.-H. and Sun, D.D., 2010. “Construction of self-organized free-standard
TiO2 nanotube arrays for effective disinfection of drinking water”. Journal of Chemical Technology and Biotechnology, Vol.
85, pp. 1061. (IF:2.045); DOI: 10.1002/jctb.2395.
Ng, W.J., 2009. “Used-water treatment and Rethinking for the future”. Hitachi Eco-Conference 2009, 16 March 2009,
Singapore.
Nguyen, Q.C. and Tan, S.K., 2009. “Quadtree mesh for combined hydrodynamic and water quality modelling”. Proceedings
of the 5th International Conference Asian and Pacific Coasts, Singapore, Vol. 2, pp. 246-251.
Nguyen, Q.C. and Tan, S.K., 2009. “Simulation of storm surge and inundation in the United States due to hurricanes using
AnuGA modelling tool”. Proceedings of the 3rd International Conference on Estuaries and Coasts, Sendai, pp. 604-609.
Nguyen, Q.C. and Tan, S.K., 2010. “Modelling of flow in Everglades National Park, Florida, USA using a quadtree grid.”
Proceedings of the 17th Congress of International Association of Hydraulics Engineering Research -- Asia Pacific Division,
Auckland, 7 pp.
Nguyen, Q.C. and Tan, S.K., 2010. “Near field mixing process of multi-port diffusers: numerical modelling with quadtree
grids”. International Symposium on Environmental Hydraulics, Athens (in press).
179
PUBLICATIONS
Nguyen, T.H.N. and Qin, X.S., 2010. “Robust optimization for water quality management under uncertainty.” Proceedings of
the 2010 Young Water Talents Symposium, Suntec, Singapore, 28 June, pp. 43-48.
Nishimura, T.*, Rahardjo, H. and Koseki, J.*, 2010. “Direct shear strength of compacted bentonite under different suctions”.
Proceedings of the 5th International Conference on Unsaturated Soils, Barcelona, Spain, 6-8 September, Vol. 1, pp. 323-328.
Nyunt, T.T., Leong, E.C. and Rahardjo, H., 2009. “Effects of matric suction and loading rate on the stiffness-strain behaviour
of kaolin”. Proceedings of the 4th Asia-Pacific Conference on Unsaturated Soils, Newcastle, Australia, 23-25 November, pp.
15-19.
O’Reilly, J.*, Lee, C., Chinalia, F.A.*, Collins, G.*, Mahony, T.* and O’Flaherty, V.*, 2010. “Microbial community dynamics
associated with biomass granulation in low-temperature (15°C) anaerobic wastewater treatment bioreactors”. Bioresource
Technology, Vol. 101, pp. 6336-6344.
O’Reilly, J.*, Lee, C., Collins, G.*, Chinalia, F.A.*, Mahony, T.* and O’Flaherty, V.*, 2009. “Quantitative and qualitative
analysis of methanogenic communities in mesophilically and psychrophilically cultivated anaerobic granular biofilms”. Water
Research, Vol. 43, pp. 3365-3374.
Ong, S.L.*, Ng, W.J. and Lee, L.Y.*, 2009. “Nitrogen removal using an anoxic-oxic ultra-compact biofilm reactor”. International
Journal of Environmental Studies, Section B, Environmental Science and Technology.
Ow, L.F., Harnas, F.R., Indrawan, I.G.B., Sahadewa, A., Sim, E.K., Rahardjo, H., Leong, E.C., Fong, Y.K. and Tan, P.Y., 2010.
“Tree pulling experiment: An analysis into the mechanical stability of rain trees”. Trees – Structure & Function (available on
line and in print).
Pan, J.H., Sun, D.D., Lee, C.M., Kim, Y.J. and Lee, W.I., 2010. “Effect of calcination temperature on the textural properties
and photocatalytic activities of highly ordered cubic mesoporous WO3/TiO2 films”. Journal of Nanoscience and Nanotechnology,
Vol. 10, No. 7, pp. 4747-4751. (IF:1.987).
Pan, T.-C., 2010. “Developing technology for multiple-hazards protection”. Proceedings of the 9th International Symposium
on New Technologies for Urban Safety of Mega Cities in Asia, 13-14 October 2010, Kobe, Japan.
Pan, T.-C., 2010. “Developing technology for protection”. Keynote Speech, Inaugural Workshop on Building Infrastructure
Protection for Homeland Security, 13 May 2010, Singapore.
Pan, T.-C., 2010. “Seismic hazard of low/moderate seismicity regions – Singapore’s perspective.” Keynote Lecture, The 10th
International ROSE School Seminar, 20-21 May 2010, EUCentre, Collegio Cardinale Riboldi, Pavia, Italy.
Pan, T.-C., Leong, C.L., Karim, R.K., Shaw, F. and Tan, A.C.T., 2009. “Explosion induced ground motion monitoring.” Final
Report No.: MINDEF-NTU/JPP/05/01, Protective Technology Research Centre, Nanyang Technological University.
Pan, T.-C, Li, B., Lu, Y.* and Lim, C.L., 2009. “Response of buildings to external and internal blast loadings.” Home Team
Journal, Singapore, Issue No. 1, pp. 67-80.
Pan, T.-C., Tan, K.H., Li, B., Fan, S.C. and Ma, G.W., 2010. “An overview of the current research programmes in Protective
Technology Research Centre at NTU.” Keynote Paper, Proceedings of the 3rd International Conference on Design and Analysis
of Protective Structures 2010 (DAPS-2010), 10-12 May 2010, Singapore, pp. K25-K39.
Peng, L., You, S.-J. and Wang, J.-Y., 2010. “Carbon nanotubes as electrode modifier promoting direct electron transfer from
Shewanella oneidensis”. Biosensors and Bioelectronics, Vol. 25, pp. 1248-1251.
Peng, L., You, S.J. and Wang, J.Y., 2010. “Electrode potential regulates cytochrome accumulation on shewanella oneidensis
cell surface and the consequence to bioelectrocatalytic current generation”. Biosensors and Bioelectronics (DDI: 10.1016/
j.bios.2010.03.039).
Prochazka, P.*, Dolezel, V.* and Lok T.S., 2009. “Optimal shape design for minimum Lagrangian”. Engineering Analysis
with Boundary Elements, Elsevier Publishing, Vol. 33, No. 4, pp. 447-455.
Qi, W., Niu, D.J. and Wang, J.-Y., 2010. “Characterization of microbial communities during hydrolysis of lignocellulosic waste
to reducing sugars”. Journal of Biobased Materials and Bioenergy. (submitted and revised one time).
Qin, X.S., Huang, G.H.* and Yu, H.*, 2009. “Enhancing remediation of LNAPL recovery through a response-surface-based
optimization approach.” Journal of Environmental Engineering (ASCE), Vol. 135, No. 10, pp. 999-1008.
180
PUBLICATIONS
Qin, X.S., 2010. “Management of environmental pollution control problems under stochastic uncertainty.” Proceedings of the
5th IEEE International Conference on Management of Innovation and Technology, Singapore, June 2-5, 2010, pp. 366-371.
Qin, X.S., 2010. “Numerical simulation of DNAPL contaminant transport and remediation in a three-dimensional heterogeneous
subsurface.” Proceedings of the 9th International Conference on Hydroinformatics, Tianjin, China, 6-10 September, pp. 566-
573.
Qin, X.S., Huang, G.H.* and He, L.*, 2010. “Development of a cluster-analysis-based distributed hydrologic modeling system.”
Proceedings of the 4th International Yellow River Forum (IYRF) on Ecological Civilization and River Ethics, Zhengzhou, China,
October 20-23, 2009, Vol. 4, pp. 46-54.
Qin, X.S., Huang, G.H.* and Liu, L.*, 2010. “A genetic-algorithm-aided chance-constrained programming model for regional
air quality management under uncertainty.” Journal of the Air & Waste Management Association, Vol. 60, No. 1, pp. 63-71.
Qin, X.S. and Xu, Y., 2010. “River water quality modeling under dual-uncertainties: A fuzzy-parameterized stochastic simulation
method.” Proceedings of the 9th International Conference on Hydroinformatics, Tianjin, China, 6-10 September, pp. 2109-
2116.
Rahardjo, H., Hua, C.J., Leong, E.C. and Santoso, V.A.*, 2010. “Performance of an Instrumented Slope under a Capillary
Barrier System”. Proceedings of the 5th International Conference on Unsaturated Soils, Barcelona, Spain, 6-8 September, Vol.
2, pp. 1279-1284.
Rahardjo, H., Leong, E.C. and Rezaur, R.B., 2009. “Laboratory characterisation of unsaturated soil for slope stability studies”.
Keynote Lecture, Proceedings of the 4th Asia-Pacific Conference on Unsaturated Soils, Newcastle, Australia, 23-25 November,
pp. 565-578.
Rahardjo, H., Ong, T.H., Rezaur, R.B., Leong, E.C. and Fredlund, D.G.*, 2010. “Response parameters for characterization of
infiltration”. Environmental Earth Sciences, Vol. 60, No. 7, pp. 1369-1380.
Rahardjo, H., Santoso, V.A.*, Leong, E.C., Ng, Y.S. and Tam, C.P.H., 2009. “Pore-water pressure characteristics of two
instrumented residual soil slopes”. Proceedings of the 4th Asia-Pacific Conference on Unsaturated Soils, Newcastle, Australia,
23-25 November, pp. 333-339.
Rahardjo, H., Satyanaga, A. Nio, Leong, E.C. and Ng, Y.S., 2010. “Effects of groundwater table position and soil properties
on stability of slope during rainfall”. ASCE Journal of Geotechnical and Geoenvironmental Engineering, November, Vol. 136,
No.11, pp. 1555-1564.
Rahardjo, H., Vilayvong, K. and Leong, E.C., 2010. “Water characteristic curves of recycled materials”. Geotechnical Testing
Journal, ASTM International, Vol. 34, No. 1, pp. 1-8 (available on line and in print).
Rahimi, A., Rahardjo, H. and Leong, E.C., 2010. “Effect of antecedent rainfall patterns on rainfall induced slope failure”.
ASCE Journal of Geotechnical and Geoenvironmental Engineering. (available on line and in print).
Rahimi, A., Rahardjo, H. and Leong, E.C., 2010. “Effect of hydraulic properties of soil on rainfall-induced slope failure”.
Journal of Engineering Geology, Vol. 114, pp. 135-143.
Reza Mohammadpour and Lim, S.Y., 2010. “Numerical modeling of three-dimensional flow around abutments in a compound
open channel”. Proceedings of 2010 International Conference on Environmental Science and Development, 26-28 February
2010, Singapore, pp. 328-332.
Santoso, V.A.*, Rahardjo, H., Leong, E.C., Ng, Y.S. and Tam, C.P.H., 2009. “Horizontal drains in residual soil slopes”.
Proceedings of the 4th Asia-Pacific Conference on Unsaturated Soils, Newcastle, Australia, 23-25 November, pp. 325-332.
Schnellmann, R.*, Busslinger, M.*, Schneider, H.* and Rahardjo, H., 2010. “Effect of rising water table in an unsaturated
slope”. Journal of Engineering Geology, Vol. 114, pp. 71-83.
Shao, D.D. and Law, A.W.K. 2009. “Salinity build-up due to brine discharges into shallow coastal waters.” Modern Physics
Letters B, Vol. 23, No. 3, pp. 541-544.
Shao, D.D. and Law, A.W.K., 2009. “Turbulent mass and momentum transport of a circular offset dense jet.” Journal of
Turbulence, Vol. 11, Issue 40, pp. 1-24.
Shao, D.D. and Law, A.W.K., 2010. “Mixing and Boundary interactions of 30 and 40 degree inclined dense jets.” Journal
of Environmental Fluid Mechanics, Vol. 10, Issue 5, pp. 521-553.
181
PUBLICATIONS
Shao, Y.B.*, Cai, Y.Q.* and Chiew, S.P., 2010. “Static strength of square tubular T-Joint under axial compression with collar
plate reinforcement”. Proceedings of the 4th International Conference on Steel and Composite Structures, Sydney, Australia,
21 – 23 July, pp. 415-421.
Shao, Y.B.*, Lie, S.T. and Chiew, S.P., 2010. “Static Strength of Tubular T-Joints with Reinforced Chord under Axial
Compression”. International Journal of Advances in Structural Engineering, Vol. 13, No. 2, pp. 369-377.
She, Q., Tang, C.Y., Wang, Y.N. and Zhang, Z., 2009. “The role of hydrodynamic conditions and solution chemistry on protein
fouling during ultrafiltration”. Desalination, Vol. 249, No. 3, pp. 1079-1087.
Shen, L. and Liu, Y., 2010. “Treatment of Ampicillin-loaded wastewater by combined adsorption and biodegradation”. Journal
of Chemical Technology and Biotechnology, Vol. 85, No. 6, pp. 814-820.
Shen, L., Liu, Y. and Paul, E.*, 2010. “A simple geometric approach for simplification of Langmuir kinetics for adsorption”.
Colloids and Surfaces A – Physical and Engineering Aspects, Vol. 349, No. 1-3, pp. 78-82.
Shi, L.*, Wang, R. and Cao, Y.*, 2009. “Effect of the rheology of poly (vinylidene fluoride-co-hexafluropropylene) (PVDF–HFP)
dope solutions on the formation of microporous hollow fibers used as membrane contactors”. Journal of Membrane Science,
Vol. 344, pp. 112-122.
Shin, S.G.*, Han, G.*, Lim, J.*, Lee, C. and Hwang, S.*, 2010. “A comprehensive microbial insight into two-stage anaerobic
digestion of food waste-recycling wastewater”. Water Research, Vol. 44, pp. 4838-4849.
Shin, S.G.*, Lee, S.*, Lee, C., Hwang, K.* and Hwang, S.*, 2010. “Qualitative and quantitative assessment of microbial
community in batch anaerobic digestion of secondary sludge”. Bioresource Technology, Vol. 101, pp. 9461-9470.
Soh, C.K. and Lim, Y.Y., 2009. “Detection and characterization of fatigue induced damage using electromechanical impedance
technique”. Multi-functional Materials and Structures II, Parts 1 & 2, Advanced Materials Research, Vol. 79-82, pp. 2031-
2034.
Su, J.*, Huang, G.H.*, Xi, B.D.* and Qin, X.S. et al., 2010. “Long-term planning of waste diversion under interval and
probabilistic uncertainties”. Resources, Conservation and Recycling, Vol. 54, No. 7, pp. 449-461.
Su, J.*, Huang, G.H.*, Xi, B.D.*, Li, Y.P.*, Qin, X.S., Huo, S.L.* and Jiang, Y.H.*, 2009. “A hybrid inexact optimization
approach for solid waste management in the City of Foshan, China.” Journal of Environmental Management, Vol. 91, No. 2,
pp. 389-402.
Sun, D.D., 2009. “Membrane water reclamation: fouling and solution”. Challenges in Environmental Science & Engineering,
CESE-2009, 14-17 July 2009, Jupiters Hotel, Townsville Australia.
Sun, D.D. and Hay Choon Teck, 2009. “Prolonged sludge retention time for high strength wastewater treatment using submerged
membrane bioreactor”. IWA Water and Industry 2009, New Zealand, 30 November to 1 December 2009.
Sun, H.Y.*, Wong, L.N.Y., Shang, Y.Q.*, Lu, Q.* and Zhan, W.*, 2010. “Systematic monitoring of the performance of anchor
systems in fractured rock masses.” International Journal of Rock Mechanics and Mining Sciences, Vol. 47, pp. 1038-1045.
Sun, J.P. and Zhao, Z.Y., 2010. “Effects of anisotropic permeability of fractured rock masses on underground oil storage
caverns.” Journal of Tunnelling and Underground Space Technology, Vol. 25, No. 5, pp. 629-637.
Sun, J.P., Zhao, Z.Y. and Zhang, Y. “Determination of three dimensional hydraulic conductivities using a combined analytical/
neural network model.” Journal of Tunnelling and Underground Space Technology, accepted.
Sun, W.*, Huang, G.H.*, Zeng, G.M.*, Qin, X.S. and Sun, X.L.*, 2009. “A stepwise-cluster microbial biomass inference
model in food waste composting.” Waste Management, Vol. 29, No. 12, pp. 2956-2968.
Talei, A., Chua, L.H.C. and Quek, C., 2010. “A novel application of a neuro-computational technique in event-based rainfall-
runoff modelling”. Expert Systems with Applications, Vol. 37, No. 12, pp. 7456-7468.
Talei, A., Chua, L.H.C. and Wong, T.S.W., 2010. “Evaluation of rainfall and discharge inputs used by Adaptive Network-Based
Fuzzy Inference Systems (ANFIS) in rainfall-runoff modelling”. Journal of Hydrology. (accepted).
Tan, C.H. (Grant), Koh, K.S.*, Rice, S.*, Zhou, Y.*, Ng, W.J. and Kjelleberg, S.*, 2010. “Concurrent quorum sensing and
quorum quenching in a simultaneous nitrification, denitrification & phosphorus removal (SNDPR) sludge community”. Biofilms4
International Conference, 1-3 September 2010, Winchester, UK.
182
PUBLICATIONS
Tan, S.B.K., Lo, E.Y.-M., Shuy, E.B., Chua, L.H.C. and Lim, W.H., 2009. “Generation of total runoff hydrographs using a
method derived from a digital filter algorithm”. Journal of Hydrologic Engineering, Vol. 14, No. 1, pp. 101-106.
Tan, S.B.K., Lo, E.Y.-M., Shuy, E.B., Chua, L.H.C. and Lim, W.H., 2009. “Hydrograph separation and development of empirical
relationships using single-parameter digital filters”. Journal of Hydrologic Engineering, Vol. 14, No. 2, pp. 271-279.
Tang, C.Y., Fu, Q.S., Gao, D., Criddle, C.S.* and Leckie, J.O.*, 2010. “Effect of solution chemistry on the adsorption of
perfluorooctane sulfonate onto mineral surfaces.” Water Research, Vol. 44, pp. 2654-2662.
Tang, C.Y., Kwon, Y.-N. and Leckie, J.O.*, 2009. “Effect of membrane chemistry and coating layer on physiochemical properties
of thin film composite polyamide RO and NF membranes. I. FTIR and XPS characterization of polyamide and coating layer
chemistry.” Desalination, Vol. 242, pp. 149-167.
Tang, C.Y., Kwon, Y.-N. and Leckie, J.O.*, 2009. “Effect of membrane chemistry and coating layer on physiochemical properties
of thin film composite polyamide RO and NF membranes. II. Membrane physiochemical properties and their dependence on
polyamide and coating layers.” Desalination, Vol. 242, pp. 168-182.
Tang, C.Y., Kwon, Y.-N. and Leckie, J.O.*, 2009. “The role of foulant-foulant electrostatic interaction on limiting flux for RO
and NF membranes during humic acid fouling-theoretical basis, experimental evidence, and AFM interaction force measurement.”
Journal of Membrane Science, Vol. 326, No. 2, pp. 526-532.
Tang, C.Y., She, Q., Lay, W.C.L., Wang, R. and Fane, A.G., 2010. “Coupled effects of internal concentration polarization and
fouling on flux behavior of forward osmosis membranes during humic acid filtration.” Journal of Membrane Science, Vol.
354, pp. 123-133.
Tang, H.W., Ding, B., Chiew, Y.M. and Fang, S.L., 2009. “Scour protection around bridge piers with tetrahedron frames”.
International Journal of Sediment Research, December, Vol. 24, No. 4, pp. 385-399.
Teo, C.C., Bhatnagar, R.* and Graves, S.C.* 2010. “Setting planned lead times for a make-to-order production system with
master schedule smoothing.” accepted for publication in IIE Transactions.
Thai, V.V., 2009. “Effective maritime security: conceptual model and empirical evidence.” Maritime Policy and Management,
Vol. 36, No. 2, pp. 147-163.
Thai, V.V., 2009. “Impact of increasing containership’s size on ports.” Proceedings of the 2nd International Symposium on
Marine Science and Technology, 2-3 November, Kaohsiung, Taiwan.
Thai, V.V., 2010. “Competency Requirements for Port Personnel in the New Era”. Proceedings of the 2010 International
Conference of Chinese Federation of Wharf Unions, 29 - 30 November, Kaohsiung, Taiwan.
Thai, V.V. and Latta, T.*, 2010. “Developing an employment brand strategy for the shortage of seafarers: The case in Australia.”
Proceedings of the 24th Australia-New Zealand Association of Management (ANZAM) Conference, 7-10 December, Adelaide,
Australia.
Thai, V.V. and Latta, T.*, 2010. “Employment brand strategy for the shortage of seafarers.” International Journal of Shipping
and Transport Logistics, Vol. 2, No. 4, pp. 411-428.
Thai, V.V., Cahoon, S.* and Tran, T.H.*, 2010. “Skill and knowledge requirements for logistics professionals in Australia.”
Proceedings of the 24th Australia-New Zealand Association of Management (ANZAM) Conference, 7-10 December, Adelaide,
Australia.
Tsai, T.T., Kao, C.M. and Wang, J.-Y., 2010. “Remediation of TCE-contaminated groundwater using Acid/BOF slag enhanced
in situ chemical oxidation”. Chemosphere (submitted).
Tunidau, J.* and Thai, V.V., 2010. “Critical factors for successful implementation of the ISM Code in some Pacific Islands
States.” WMU Journal of Maritime Affairs, Vol. 9, No. 1, pp. 63-80.
Vu, T.-T. and Tan, S.-K., 2009. “A review of the current state-of-the-arts on the application of silt screens as sediment control
equipment in open water”. Proceedings of the 5th International Conference on Asian and Pacific Coasts, Singapore, 13-16
October 2009, Vol. 2, pp. 60-66.
Vu, T.-T. and Tan, S.-K., 2010. “Laboratory investigation of hydraulic performance of silt screen”. Proceedings of the 9th
International Conference on Hydrodynamics, Shanghai, China, October 2010 (in press).
183
PUBLICATIONS
Vu, T.-T., Tan, S.-K., and Stéphanie Doorn-Groen, 2010. “A case study of silt screen deployment”. World Dredging Congress
XIX, Beijing, China, September 2010 (in press).
Vu, T.-T. and Tan, S.-K., 2010. “Laboratory investigation of hydraulic performance of silt screen”. Proceedings of the 9th
International Conference on Hydrodynamics, Shanghai, China, October 2010 (in press).
Vu, T.-T., Tan, S.-K., and Stéphanie Doorn-Groen, 2010. “A case study of silt screen deployment”. World Dredging Congress
XIX, Beijing, China, September 2010 (in press).
Wanatowski, D. *and Chu, J., 2009. “Instability behavior of Changi sand in plane-strain tests.” Proceedings of the 17th
International Conference on Soil Mechanics and Geotechnical Engineering, Alexandria, Egypt, 5-9 October, Vol. 1, pp. 89-
92.
Wanatowski, D.* and Chu, J., 2010. Discussion of “Unstable behaviour of model Jamuna micaceous sand”. Geotechnique,
Vol. 60, No. 4, pp. 307-308.
Wanatowski, D.*, Chu, J. and Gan, C-L.*, 2009. “Compressibility of Changi sand in Ko consolidation.” Geomechanics and
Engineering: An International Journal, Vol. 1, No. 3, pp. 241-258.
Wanatowski, D.*, Chu, J. and Lo, S-C.*, 2010. “New types of failure mechanisms for flowslide.” Geomechanics and
Geoengineering: An International Journal, Vol. 5, No. 1, pp. 3-13.
Wanatowski, D.*, Chu, J. and Loke, W.L.*, 2010. “Drained instability of sand in plane strain.” Canadian Geotechnical
Journal, Vol. 47, No. 4, pp. 400-412.
Wang, R., Lei, S., Tang, C.Y., Chou, S., Qiu, C. and Fane, A.G., 2010. “Characterization of novel forward osmosis hollow
fiber membranes.” Journal of Membrane Science, Vol. 355, pp. 158-167.
Wang, R., Shi, L.*, Tang, C.Y., Chou, S., Qiu, C. and Fane, A.G., 2010. “Characterization of novel forward osmosis hollow
fiber membranes”. Journal of Membrane Science, Vol. 355, pp. 158-167.
Wang, R.Q.*, Law, A.W.K., Adams, E.E.* and Fringer, O.B.*, 2009. “The determination of formation number for starting
buoyant jet”. Proceedings of the 2nd International Symposium on Computational Mechanics (ISCM II) and 12th International
Conference on Enhancement and Promotion of Computational Methods in Engineering and Science (EPMESC XII), 30 November
– 3 December, Hong Kong – Macau.
Wang, R.Q.*, Law, A.W.K., Adams, E.E.* and Fringer, O.B.*, 2009. “Large-eddy simulation of starting buoyant jets.”
Proceedings of the 33rd IAHR Congress, 9-14 August, Vancouver, Canada.
Wang, R.Q., Law, A.W.K., Adams, E.E.* and Fringer, O.B.*, 2009. “Buoyant formation number of a starting buoyant jet.”
Physics of Fluids, Vol. 21, Issue 12, Article No. 125114.
Wang, X.K., Hao, Z. and Tan, S.K., 2010. “Flow around a pipeline near a smooth bed in steady current”. Proceedings
of the 29th International Conference on Ocean, Offshore and Arctic Engineering (OMAE2010), 6-11 June, Shanghai, China,
OMAE2010-20749.
Wang, X.K. and Tan, S.K., 2010. “Lagrangian modelling of tidal bores passing through bridge piers”. Proceedings of the 9th
International Conference on Hydrodynamics (ICHD2010) (in press).
Wang, Y.J., Xu, S.P., Zhang, X.W. and Sun, D.D., 2010. “Enhanced photocatalytic oxidation by anatase nanotubes for water
treatment”. IWA LET Conference, USA, 3 June 2010.
Wei, X.Y., Zhao, Z.Y. and Gu, J., 2009. “Numerical simulations of rock mass damage induced by underground explosion.”
International Journal of Rock Mechanics and Mining Sciences, Vol. 46, No. 7, pp. 1206-1213.
Wicaksana, F., Fane, A.G.F. and Law, A.W.K., 2009. “The use of constant temperature anemometry for permeate flow distribution
measurement in a submerged hollow fibre system.” Journal of Membrane Science, Vol. 339, Issues 1-2, pp. 195-203.
Wong, L.N.Y. and Einstein, H.H.*, 2009. “Crack coalescence in molded gypsum and carrara marble: Part 1 – Macroscopic
observations and interpretation.” Rock Mechanics and Rock Engineering, Vol. 42, No. 3, pp. 475-511.
Wong L.N.Y. and Einstein, H.H.*, 2009. “Crack coalescence in molded gypsum and carrara marble: Part 2 – Microscopic
observations and interpretation.” Rock Mechanics and Rock Engineering, Vol. 42, No. 3, pp. 513-545.
184
PUBLICATIONS
Wong, L.N.Y. and Einstein, H.H.*, 2009. “Process zone development associated with cracking processes in carrara marble.”
Proceedings of the 9th International Conference on Analysis of Discontinuous Deformation - New Developments and Applications,
Nanyang Technological University, Singapore, 25-27 November, pp. 581-588.
Wong, L.N.Y. and Einstein, H.H.*, 2009. “Systematic evaluation of cracking behavior in specimens containing single flaws
under uniaxial compression.” International Journal of Rock Mechanics and Mining Sciences, Vol. 46, No. 2, pp. 239-249.
Wong, L.N.Y. and Einstein, H.H.*, 2009. “Using high speed video imaging technique in the study of cracking and coalescence
processes in rock.” Geotechnical Testing Journal, Vol. 32, No. 2, pp. 164-180.
Wong, L.N.Y. and Liu, G., 2010. “An improved K-means clustering method for the automatic grouping of discontinuity sets.”
Proceedings of the 44th U.S. Rock Mechanics Symposium and 5th U.S.-Canada Rock Mechanics Symposium, Salt Lake City,
UT. ARMA 10-265.
Wu, T. and Li, B., 2009. “Experimental verification of continuous deep beams with openings designed using strut-and-tie
modelling”. IES Journal Part A: Civil and Structural Engineering, Vol. 2, No. 4, November, pp. 282–295.
Wu, Y.S. and Chiew, Y.M., 2009. “Experimental study on 3-Dimensional scour at submarine pipelines”. Proceedings of the
33rd IAHR Congress, 9-14 August, Vancouver, British Columbia, Canada, pp. 2198-2206.
Xi, B.D.*, Su, J.*, Huang, G.H.* and Qin, X.S., et al. 2010. “An integrated optimization approach and multi-criteria decision
analysis for supporting the waste-management system of the City of Beijing, China.” Engineering Applications of Artificial
Intelligence, Vol. 23, No. 4, pp. 620-631.
Xiong, Y.H. and Liu, Y., 2010. “Involvement of ATP and Autoinducer-2 in Aerobic Granulation”. Biotechnology and
Bioengineering, Vol. 105, No. 1, pp. 51-58.
Xiong, Y.H. and Liu, Y., 2010. “Biological control of microbial attachment: a promising alternative for mitigating membrane
biofouling”. Applied Microbiology and Biotechnology, Vol. 86, No. 3, pp. 825-837.
Xu, H.J. and Liu, Y., 2010. “Control of microbial attachment by inhibition of ATP and ATP-Mediated Autoinducer-2”.
Biotechnology and Bioengineering, Vol. 107, No. 1, pp. 31-36.
Xu, S.P. and Sun, D.D., 2009. “Significant improvement of hydrogen generation rate using TiO2 photocatalyst enhanced with
Cu”. International Journal of Hydrogen Energy, Vol. 34, Issue 15, pp. 6096-6104. (IF: 3.452).
Xu, S.P., Ng, J.W., Zhang, X.W., Bai, H.W. and Sun, D.D., 2010. “Fabrication and comparison of highly-efficient Cu incorporated
TiO2 photocatalyst for hydrogen generation from water”. International Journal of Hydrogen Energy, Vol. 35, pp. 5254-5261.
(IF:3.945).
Xu, S.P., Ng, J.W., Zhang, X.W., Bai, H.W. and Sun, D.D., 2010. “Adsorption and photocatalytic degradation of dye pollutants
over TiO2 nanotube photocatalyst”. Proceedings of the 6th International Conference on Interfaces Against Pollution, 16-19 May
2010, Beijing, China.
Xu, Y., Huang, G.H.* and Qin, X.S., 2009. “An inexact two-stage stochastic robust optimization model for water resources
management under uncertainty.” Environmental Engineering Science, Vol. 26, No. 12, pp. 1765-1767.
Xu, Y., Huang, G.H.* and Qin, X.S., 2010. “An inexact fuzzy-chance-constrained air quality management model.” Journal
of the Air and Waste Management Association, Vol. 60, No. 7, pp. 805-819.
Xu, Y., Huang, G.H.*, Qin, X.S. and Cao, M.F.*, 2009. “SRCCP: A stochastic robust chance-constrained programming model for
municipal solid waste management under uncertainty.” Resources, Conservation and Recycling, Vol. 53, No. 6, pp. 352-363.
Xu, Y., Huang, G.H.*, Qin, X.S. and Huang, Y., 2009. “SRFILP: A stochastic robust fuzzy interval linear programming model
for municipal solid waste management under uncertainty.” Journal of Environmental Informatics, Vol. 14, No. 2, pp. 1-9.
Xu, Y., Huang, G.H.*, Qin, X.S., Cao, M.F.* and Sun, Y.*, 2009. “An interval-parameter robust optimization model for
supporting municipal solid waste management under uncertainty.” Waste Management, Vol. 30, No. 2, pp. 316-327.
Xu, Y., Peng, X., Tang, C.Y., Fu, Q.S. and Nie, S., 2010. “Effect of draw solution concentration and operating conditions on
forward osmosis and pressure retarded osmosis performance in a spiral wound module.” Journal of Membrane Science, Vol.
348, No. 1-2, pp. 298-309.
185
PUBLICATIONS
Yan, S.W.* and Chu, J., 2010. “Construction of an offshore dike using slurry filled geotextile mats.” Geotextiles and
Geomembranes (available online on 14 Feb 2010).
Yang, A.L.*, Huang, G.H.* and Qin, X.S., 2010. “An integrated simulation-assessment approach for evaluating health risks of
groundwater contamination under multiple uncertainties.” Water Resources Management, Vol. 24, No.13, pp. 3349-3369.
Yang, H., Rahardjo, H. and Xiao, D., 2010. “Rapid drawdown of water table in layered soil column.” ASCE Geotechnical
Special Publication No. 204, Geoenvironmental Engineering and Geotechnics, Proceedings of GeoShanghai International
Conference, 3-5 June, ed. by He, Q. and Shen, S.L., pp. 202-209.
Yang, H.Y., Yu, S.F., Lau, S.P., Zhang, X., Sun, D.D. and Jun, G., 2009. “Direct growth of ZnO nanocrystals onto the surface
of porous TiO2 nanotube arrays for highly efficient and recyclable photocatalysts”. Small, Vol. 5, Issue 20, pp. 2260-2264.
(IF: 6.525).
Yoo, K.S. and Rahardjo, H., 2009. “Performance of modified volumetric pressure plate”. Proceedings of the 4th Asia-Pacific
Conference on Unsaturated Soils, Newcastle, Australia, 23-25 November, pp. 191-196.
Yoo, K.S. and Rahardjo, H., 2010. “Modification of volumetric pressure plate extractor“. Journal of ASTM International,
October 2010, Vol. 7, No. 9.
You, S.J., Ren, N.Q., Wang, J.Y., Zhao, Q.L., Yang, F.L., Zhang, J.N., Fu, L. and Luo, P., 2009. “Improving phosphate buffer
free cathode performance of microbial fuel cell based on biological nitrification”. Biosensors and Bioelectronics (accepted).
You, S.J., Ren, N.Q., Zhao, Q.L., Wang, J.Y. and Yang, F.L., 2009. “Power generation and electrochemical analysis of biocathode
microbial fuel cell using graphite fibre brush as cathode material”. Fuel Cells, Vol. 9, pp. 588-596.
You, S.J., Wang, J.Y., Ren, N.Q., Wang, X.H. and Zhang, J.N., 2010. “Sustainable conversion of glucose into hydrogen peroxide
in a solid polymer electrolyte microbial fuel cell”. ChemSusChem, Vol. 3, pp. 334-338.
You, S.J., Wang, X.H., Zhang, J.N., Wang, J.Y., Ren, N.Q. and Gong, X.B., 2010. “Fabrication of stainless steel mesh gas
diffusion electrode for power generation in microbial fuel cell”. Biosensors and Bioelectronics, Vol. 25, No. 5, pp. 1248-
1251.
Yu Yao, Huang, Z.H., Lo, E.Y.M. and Monismith, S.G.*, 2009. “Wave-induced set-up over fringing coral reefs: A comparison
between smooth and porous reef flats”. Proceedings of the 5th International Conference on Asian and Pacific Coasts, Singapore,
World Scientific Publishing, pp. 236-244.
Yuan, Z.D. and Huang, Z.H., 2009. “Solitary wave forces on an array of closely spaced circular cylinders”. Proceedings of
the 5th International Conference on Asian and Pacific Coasts, Singapore, World Scientific Publishing, pp. 136-142.
Zhang, X.W., Zhang, T., Ng, J.W., Pan, J.H. and Sun, D.D., 2010. “Transformation of bromine species in TiO2 photocatalytic
system”. Environmental Science & Technology, Vol. 44, No. 1, pp. 439-444. (IF: 4.458).
Zhang, Bo, Liu, H.L.* and Chu, J., 2009. “Analysis of stress distribution in large diameter concrete pipe pile supported
embankment.” Proceedings of the International Symposium on Ground Improvement Technologies and Case Histories, Eds,
Leung, C.F., Chu, J., and Shen, R.F., Research Publishing, pp. 239-245.
Zhang, D.Q., Tan, S.K. and Gersberg, R.M.*, 2010. “Gray water treatment in an Urban Area of Beijing, China”. Proceedings
of the 4th International Conference on Bioinformatics and Biomedical Engineering, 21-23 June 2010, Chengdu, China (in
press).
Zhang, D.Q., Gersberg, R.M.* and Tan, S.K. 2009. “Constructed wetlands in China”. Journal of Ecological Engineering,
http://dx.doi.org/10.1016/_j.ecoleng.2009.07.007, Vol. 35, pp. 1367-1378.
Zhang, D.Q., Tan, S.K. and Gersberg, R.M.*, 2010. “A comparison of municipal solid waste management in Berlin and
Singapore”. Waste Management, Elsevier, Vol. 30, pp. 921-933.
Zhang, D.Q., Tan, S.K. and Gersberg, R.M.*, 2010. “Urban solid waste management in China: Status, problems and challenges”.
Journal of Environmental Management, Vol. 91, pp. 1623-33.
Zhang, D.Q., Gersberg, Richard* M., Wilhelm, C.* and Voigt, M.*, 2010. “Decentralized wastewater management in an urban
area of Beijing, China – rainwater harvesting and greywater reuse”. Urban Water Journal, Vol. 6, No. 5, pp. 375-385.
186
PUBLICATIONS
Zhang, J.S., Zhou, J.*, Liu, Y. and Fane, A.G., 2010. “A comparison of membrane fouling under constant and variable organic
loadings in submerge membrane bioreactors”. Water Research, Vol. 44, No. 18, pp. 5407-5413.
Zhang, L.L.*, He, D.*, Chen, J.M.* and Liu, Y., 2010. “Biodegradation of 2-chloroaniline, 3-chloroaniline, and 4-chloroaniline
by a novel strain Delftia tsuruhatensis HI”. Journal of Hazardous Materials, Vol. 179, No. 1-3, pp. 875-882.
Zhang, Q.W., Megawati, K., Huang, L.P. and Pan, T.C., 2010. “A new static nonlinear procedure for assessment of seismic
performance of high-rise buildings.” Proceedings of the 8th International Conference on Fracture and Strength of Solids, Kuala
Lumpur, 7-9 June 2010.
Zhang, X.W., Pan, J.H., Du, A.J., Xu, S.P. and Sun, D.D., 2009. “Room-temperature fabrication of anatase TiO2 submicrospheres
with nanothornlike shell for photocatalytic degradation of methylene blue”. Journal of Photochemistry and Photobiology A:
Chemistry, Vol. 204, Issue 2-3, pp. 154-160. (IF:20657).
Zhang, X.W., Zhang, T., Ng, J.W. and Sun, D.D., 2009. “High-performance multifunctional TiO2 nanowire ultrafiltration membrane
with hierarchical layer structure for water treatment”. Advanced Functional Materials, Vol. 19, pp. 3731-3736. (IF: 6.99).
Zhang, X.W., Zhang, T., Ng, J.W. and Sun, D.D., 2010. “Superhydrophilic manganese oxide nanowire membrane for high flux
water filtration”. Singapore International Water Week, Singapore, 28 June - 3 July 2010.
Zhang, Y.P., Fane, A.G. and Law, A.W.K., 2010. “Critical flux and particle deposition of fractal flocs during crossflow
microfiltration.” Journal of Membrane Science, Vol. 353, Issues 1-2, pp. 28-35.
Zhao, Z.Y. and Bao, H.R., 2010. “Resolving the indeterminacy of the vertex-vertex contact in the 2D discontinuous deformation
analysis.” Proceedings of the 5th International Conference on Discrete Element Methods, London, pp. 86-91.
Zhao, Z.Y. and Gu, J., 2009. “Stress recovery procedure for discontinuous deformation analysis.” Journal of Advances in
Engineering Software, Vol. 40, No. 1, pp. 52-57.
Zhao, Z.Y., Zhang, Y. and Wei, X.Y., 2009. “Discontinuous deformation analysis for parallel hole cut blasting in rock mass.”
Proceedings of the 9th International Conference on Analysis of Discontinuous Deformations (ICADD-9), Singapore, pp. 169-
176.
Zheng, H., Megawati, K. and Yokoi, I.*, 2010. “Development of user-friendly application software for structural health
monitoring.” Proceedings of the 5th World Conference on Structural Control and Monitoring, Tokyo, 12-14 July 2010.
Zhou, Q.* and Cheng, N.S., 2009. “Experimental investigation of single particle settling in turbulence generated by oscillating
grid.” Chemical Engineering Journal, Vol. 149, No. 1-3, pp. 289-300.
Zhou, Y.*, Ganda, L., Lim, M., Yuan, Z.*, Kjelleberg, S.* and Ng, W.J., 2010. “Free nitrous acid (FNA) inhibition on denitrifying
poly-phosphate accumulating organisms (DPAOs)”. Applied Microbiology and Biotechnology, Vol. 88, No. 1, pp. 359-369.
Zhou, Y.*, Harjono, S., Lim, M. and Ng, W.J., 2010. “Nitrous oxide (N2O) accumulation by denitrifying polyphosphate
accumulating organism (DPAO) using internal carbon source”. Water and Energy, 10-12 November, Amsterdam, Netherlands.
187
CONTENTS
MESSAGE FROM CHAIR 1 • Technologies for Water Softening: a review 80

• Time-sequence analysis of jet-flipping of localized
SUSTAINABLE URBAN SYSTEMS 1 scour by 2-D wall jets 84
CEE VISION AND MISSION 6 INFRASTRUCTURE SYSTEMS AND MARITIME
STATISTICS 7 STUDIES
• Faculty & Staff 7 • A decision support system for port selection 88
• Publications, patents and research grants 7 • Dangerous goods regulating system in Singapore 90
• Student Enrolment 7 • Determination of Coefficient of Consolidation by
Rowe Cell 94
UNDERGRADUATE PROGRAMMES 8
• Effects of electric vehicles on climate goals – Singapore
GRADUATE PROGRAMMES 10 and Germany in comparison 96
ACHIEVEMENTS AND COMMENDATIONS 11 • Maritime Studies Degree Programmes in Shipping
RESEARCH CENTRES Management – An International Comparison 100
• Centre for Infrastructure Systems (CIS) 13 • P-wave velocity measurements in sedimentary rocks 104
• DHI-NTU Centre 16 • Role of filler in macro structure of asphalt mixture
and its binding characteristic with asphalt 108
• Maritime Research Centre (MRC) 19
• Protective Technology Research Centre (PTRC) 21 STRUCTURES AND MECHANICS
• Residues and Resource Reclamation Centre (R3C) 25 • A co-rotational shell element with material
nonlinearities 113
ENVIRONMENTAL AND WATER RESOURCES • Collision analysis of offshore flexible risers 117
ENGINEERING • Effects of anisotropic permeability of fractured rock
• A laboratory study of wave-induced setup over coral masses on rock caverns 120
reefs with an idealized ridge 28 • Entropy based ensemble neural network design 124
• Adoption and acceptance of CNG vehicles on • Burst strength estimation of a cracked compressed
the urban environment 31 natural gas (CNG) tank cylinder 127
• Adsorption thermodynamics of antibiotics by GAC 35 • Experimental Study and numerical modeling
• An interval approach for supporting urban of stress concentration factor in high strength steel
water supply analysis 38 plate-to-plate Y joints 130
• Anaerobic hydrolysis of particulates in sewage 41 • Experimental tests of different types of steel
• Data-driven approach for multi-step ahead beam-column joints subjected to catenary action 134
flood forecasting for the Lower Mekong 45 • Modeling of piezoelectric energy harvester 138
• Design of brine outfall for seawater reverse • Numerical simulation of steel bolted beam-column
osmosis (SWRO) desalination plants 51 connections subjected to dynamic loading 142
• Fouling behavior of forward osmosis membranes 53 • Experiment investigation on residual stress distributions
• Estimate of resistance induced by simulated of high strength steel plate-to-plate Y joints 146
emergent vegetation in open channel flows 56
• Robustness of steel angle beam-column joints
• Life cycle analysis of offshore gangway 59 under column removal scenarios 150
• Dam-break flow simulation with sediment transport 62 • Seismic responses of reinforced concrete buildings
• Numerical simulation of wedge water entry based on with wall-like columns 154
two-phase SPH model 65 • Consistency of shear-wave velocity structures
• Optimization and enhancement of microbial hydrolysis inferred from microtremor observations 157
of lignocellulosic waste to reducing sugars 68
RESEARCH PROJECTS
• Removal of pharmaceutical compounds in tropical
• Ongoing projects 161
constructed wetlands 71
• Completed projects 164
• Responses of floating breakwater to regular waves 75
• Solitary wave interaction with a slotted barrier: • PhD Theses 166
wave scattering and hydrodynamic forces 77 PUBLICATIONS 171
EDITORIAL BOARD ADDITIONAL COPIES AND ENQUIRIES

Leong Eng Choon – Chairman For general enquiries about this publication and
Cheng Nian Sheng request for additional copies, please write to:
Jim Chen
Low Ying Min Chair
Teh Cee Ing School of Civil and Environmental Engineering
Nanyang Technological University
Tiong Lee Kong, Robert
50 Nanyang Avenue
Singapore 639798
EDITORIAL ASSISTANT Tel: 65-67905264
Fax: 65-67910676
Jamillah Bte Sa’adon
Email: D-CEE@ntu.edu.sg
Published by School of Civil and Environmental Engineering, Nanyang Technological University, Singapore Printed by PHOTOPLATES PTE LTD

RB 24

Uploaded by

Copyright:

Available Formats

RB 24

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

RB 24

Uploaded by

Copyright:

Available Formats

CIVIL ENGINEERING

RESEARCH ISSN 0219-0370 • No. 24 / 2011 • www.ntu.edu.sg/cee/research/bulletin/index.asp

SUSTAINABLE URBAN SYSTEMS

MESSAGE FROM THE CHAIR

O ur School of Civil and

This year also saw CEE embarking on two new research

Message from the Chair (Cont’d from Pg 1)

Sustainable Urban Systems (Cont’d from Pg 1)

The second project is titled “Sustainable Urban Waste

The following provides a synopsis of the two CRPs.

caverns and offshore reclamation for space creation. Both

The expected outcome of this research is the establishment

Collaborators: RESEARCH OUTCOME

• A technology for phosphorous - and nitrogen-recovery KEY MEMBERS OF PROGRAM TEAM:

• A technology for non-invasive geophysical site

CEE VISION & MISSION

practice of civil and environmental engineering

Publications, Patents and Research Grant

Year 2006 2007 2008 2009 2010

Programme/ Under- MEng PhD PhD MSc MSc MSc MSc

AY2008-09 1210 29 163 6 216 93 98 256

BACHELOR DEGREE PROGRAMMES

in a private company or public organization, where they

Civil Engineering Research • January 2011

Engineering) Sediment Transport

Civil Engineering Research • January 2011

(1) Professor Sun’s Research on TiO2 solar cell was

INVITED LECTURES • Invited Lecture: “Recent development in the

Associate Professor Chu Jian was invited to deliver

Change including Global Warming, 3-4 December

Activities of Centre for Infrastructure Systems (CIS)

4. Invited Public Lecture on “Reinforced Soil System:

(B) Research and Development Research Project on Study of Transport Energy

2) Kyoto Seminar 2010 on Geotechnics/Earthquake

Civil Engineering Research • January 2011

Activities of DHI-NTU Centre in 2010

COLLABORATIONS AND c) National Research Fund (NRF): on hydraulics

Activities of Maritime Research Centre (MRC)

investigation for safety assessment of cracked square

“Hydrodynamics of trapezoidal embankment weirs”. Design of Container Bund, October-November, 2010

Activities of Protective Technology Research Centre

underground rock cavern projects

Visit by Dr Lim Chee Onn, NTU Board of

Third International Conference on Design and Analysis of

INTERNATIONAL AND LOCAL

Visit by Delegates from the MINDEF, 26 April

* The delegation was led by BG(NS) Ravinder Singh,

Visit by Delegates from MINDEF

Residues and Resource Reclamation Centre (R3C)

Converting wastes into new and useful materials (e.g. plastic

• Acting as a think tank to identify R3 research needs

R3 Resource and Technology Transfer

Education and Training for R3 Professionals

• Providing research scholarships for PhD programme

CENTRE ACTIVITIES • Site selection for sanitary landfill using GIS

• Production of biofuels and biochemicals from

• Use of molecular tools for quantifying slowly

• Arsenic decontamination of groundwater

Workshop News Updates