Michel Chaaya
Development and Integration of a Visual Design Management System for a Life Cycle Project Management Model
Supervisor: Prof Ali Jaafari

This thesis presents the synthesis of a design management system within a life cycle project management (LCPM) model. The LCPM model is based on:
· integrating all project phases into a single phase, using integrated teams/decision approaches and incorporating concurrency concepts, as well as an advance alliance method;
· developing a consistent and efficient platform for development of a candidate project from concept to completion and through to facility operation;
· integrating project information across project life cycle;
· analysing holistically project variables and executing project management functions, using business objectives as the basis for on-going decision-making and evaluation;
· promoting interdisciplinary communication and teamwork; and
· integrating the processes for planning, engineering, documentation, procurement, and construction management throughout the project life cycle.

The thesis focuses on the integration of design information within the LCPM model to support collaborative decision-making in a concurrent engineering/construction environment. For this purpose, a visual design management system (VDM) was developed to assist and support the management of design information. VDM is an integrated extension to the original computer program, known as SPMIS (Smart Project Management Information System) which has been designed and developed to incorporate the philosophies adopted by the LCPM model. VDM (like SPMIS) is a Windows-based application that utilises a database management system and the object-oriented modelling techniques. The VDM will establish a set of protocols for design information input and retrieval and is integrated with SPMIS. It will assist the project management team in evaluating the teams' inputs in real time against the life cycle objective functions established at the outset.

The thesis also highlights the implementation of integrated design management processes in the LCPM model, and specifically those related to generating and integrating design in real time basis, as opposed to discipline-based individual designs superimposed onto one another. Since under the LCPM approach projects are treated as business ventures, the client is no longer the sole party who can benefit from the success of the project's business. The project must be designed and optimised in such a manner so as to strike a fair balance between the interests of the stakeholders and the financial gain by the client body. The definition and broad design of the facility and its components will be a collective responsibility, discharged by composite teams whose members are drawn from the respective participants.

The thesis presents the results of a survey to establish the design management processes as applied to real projects and to give an overview of the extent of the application of Information Technology in such processes. The survey results have assisted the author in synthesising the proposed design management processes in the LCPM model.

The thesis also presents VIRCON (VIRtual CONstruction) which incorporates a visualisation system and a construction management information system (CMIS). The author contributed to the research and development of the visualisation system which was considered a test-bed to assist the author in integrating and testing various object models, behaviours and tools specially those related to simulating CAD information (2D and 3D models). The validation of VIRCON was to explore if the objectives behind developing and integrating the VDM system within SPMIS were attainable.

Vijay Gupta
Micro Zonation of Delhi (India)
Supervisor: Prof Harry Poulos

Delhi, the capital of India, is a fast developing city located in a highly earthquake-prone belt near the world famous seismically active Himalayan region. The heavily populated city with high rise structures could be prone to damage under the occurrence of a natural catastrophe such as an earthquake. Recent earthquakes of magnitude greater than 5.0 in and around Delhi have necessitated a micro-zonation study which can represent the anticipated behaviour of structures in different parts of Delhi under seismic action.

This study presents a summary of the geology, characteristic soil profiles and representative earthquake profiles collected from various departments, organisations and institutions in India. An approach adopted towards the micro-zoning study for Delhi region is outlined with respect to the following issues: (i) characterization of typical soil profiles, (ii) characterization of dynamic soil properties, (iii) characterization of representative earthquakes, (iv) analysis of amplification of earthquake motions and (v) assessment of liquefaction potential. The effects on bedrock earthquake spectra of the soil characteristics present over the bedrock, structure stiffness etc. have been studied and results obtained are presented.

Micro-zonation by numerical modelling of seismic wave amplification/attenuation in different soil profiles has been carried out for several collected earthquake time-acceleration records at the base of the layer. The geotechnical properties including dynamic characteristics of soil layers in a profile have been deduced from boring and testing results and correlation curves. The spectra obtained at the surface from a numerical analysis have been used to develop proposed modifications of the Indian code average acceleration spectral curves. A compendium of these modified spectral curves provides the peak spectral characteristics corresponding to different structural natural periods for various cells of the 3 km square grid into which Delhi has been divided for the current study. The study also outlines the further research needed to obtain more accurate results.

Mohammed Kabirul Islam
Constitutive Models for Carbonate Sand and their Application to Footing Problems
Supervisor: Prof John Carter

A numerical study was carried out to model the stress-strain or constitutive behaviour of carbonate sands. Several existing critical state models were used to simulate the triaxial response of both cemented and uncemented carbonate sands under drained conditions. Two new models were proposed as relatively simple modifications of the Cam Clay and Modified Cam Clay model respectively. It was observed that the proposed as well as the existing critical state models, which correctly take into account the yield, plastic stress-dilatancy and volumetric compression behaviour of carbonate sands, can predict reasonably well their triaxial response. The effect of cementation was simulated in the proposed models by the increased preconsolidation pressure occurring as a result of cementation. The triaxial behaviour of dense artificially cemented carbonate sands could be predicted quite well using this approach. However, the stress-strain behaviour of loose cemented carbonates was not predicted very well using any of the existing critical state models including the proposed ones.

All of these critical state models were implemented in a finite element procedure. The models were then used to simulate the drained pressure-displacement response of circular footings resting on artificially cemented carbonate sand. It was observed that some of the models could predict quite well, both qualitatively and quantitatively, the drained pressure-displacement response of circular footings resting on cemented carbonate sand.

One of the proposed models (called here the SU2 model) was used to predict the centrifuge pressure-displacement data of surface circular footings as well as of model-scale footings embedded in carbonate sand. It was observed that the proposed model gives reasonable predictions of footing behaviour in both cases. A parametric study of the pressure-displacement behaviour of surface circular footings and of footings embedded in carbonate sand was carried out using the model. The effect of model parameters on footing response was investigated and simplified design charts were proposed for an assumed idealised soil profile below the footing. The proposed design charts may be used to obtain approximate bilinear pressure curves for circular footings on carbonate sand subjected to vertical load. Separate design charts were proposed for circular footings resting on normally consolidated carbonate sand.

Finally, the pressure-displacement response of circular footings subjected to inclined load was simulated using the proposed SU2 model in a semi-analytical three dimensional finite element procedure. Satisfactory predictions of the pressure-displacement behaviour of the experimental model-scale footing resting on cemented carbonate sand and subjected to inclined load could be made using the model. However, similar predictions for uncemented sand was not satisfactory. A simplified footing model was proposed to predict the pressure-displacement response of circular footings subjected to inclined load without the use of a 3D finite element procedure. The simplified model could simulate quite well the response for footings obtained using the 3D finite element procedure.

Takashi Itakura
Transport of organic contaminants through natural clayey soils
Supervisor: Dr. D. W. Airey
Associate Supervisor: Prof. J. P. Carter

The thesis characterises the general and fundamental transport of organic contaminants through soils at the Castlereagh waste management centre located 50km to the west of Sydney, NSW. At the site a natural clay deposit called the Londonderry clay is used to contain a range of liquid and sludge industrial wastes. The Londonderry clay is underlain by the Rickabys Creek gravel and the Bringelly shale. Characterisation of contaminant transport has been based on the advective and diffusive transport model that includes the effects of advection, sorption, diffusion and dispersion and biodegradation. Tests have been performed to characterise the three soil layers and to determine parameters that describe the advective-diffusive model.

The deposition environment of the Londonderry clay and the Rickabys Creek gravel has been influenced by floodplains associated with the Hawkesbury-Nepean Catchment. These two layers have various soil particle size distributions, an indication of the heterogeneity of the Tertiary sequence, and generally significant clay contents. The clay minerals are predominantly kaolinitic with the co-presence of illite. The Bringelly shale consists of various sedimentary rocks and the clay mineralogy is dependent on the rock type. Claystone members contain a significant proportion of swelling clay minerals such as montmorillonite while non-swelling clay minerals such as kaolinite and illite are the major clay minerals for sandstone members.

To assess the advective transport, the saturated hydraulic conductivities (k) of the soils at the site have been measured using the triaxial compression testing apparatus. The results have been used to construct relationships between stress, void ratio and k for these soils. The Londonderry clay and the Rickabys Creek gravel have a wide range of k values, from 10^-11 to 10^-7 m/s, because of the heterogeneity of the layers. The Bringelly shale has k values of the order of 10-12 to 10-11 m/s.

Three organic compounds, methyl ethyl ketone (MEK), toluene and trichloroethylene (TCE) were used to determine the distribution coefficients (K_d) of these compounds for the Londonderry clay and the Rickabys Creek gravel using the batch type sorption testing method. The Kd values suggest that MEK is not sorbed to the soils while TCE and toluene are weakly sorbed. The K_d values determined for TCE and toluene range from 1.09 to 1.93 l/kg and from 1.81 to 2.95 l/kg, respectively. The K_d values have been found to be unaffected by the electrical conductivity, the equilibrium time, the presence of other organic compounds. These values have also been found to be greater than the organic carbon fraction based K_d estimation method.

The effective diffusion coefficients (D_e) and K_d values of the above three compounds have been measured from diffusion tests. These two coefficients have not been significantly affected by porosity, sample thickness, concentration and the presence of other compounds. The D_e values determined range approximately from 1 x 10^-6 to 3 x 10^-6 cm²/s. The K_d values estimated are greater than expected from the organic carbon fraction, but smaller than those from the batch type sorption tests.

No microbiological activity, and only a small microbial biomass, has been found in the Londonderry clay the sample tested. The biodegradation of seven organic compounds (benzene, ethylbenzene, toluene, m-xylene, o-xylene, MEK and TCE) has been measured under a denitrifying anaerobic condition similar to that in situ. No biodegradation was observed, however, with addition of nitrogen and phosphorus nutrients, benzene, MEK and TCE were weakly biodegraded and the results have been used to determine first order bio-decay constants for these three compounds.

The long-term performance of the Londonderry clay has been simulated numerically using the experimentally determined transport parameters. It has been found that even when it is assumed that the Londonderry clay is homogeneously structured, the clay cannot securely contain hydrophilic organic contaminants such as MEK. The more hydrophobic organic contaminants, TCE and toluene, also migrate through the clay layer, but at a slower rate because of their greater K_d values. The presence of sand patches in the Londonderry clay has been found to significantly influence contaminant transport, allowing the contaminants to migrate faster particularly in the vertical direction.

Nagendran Loganathan
Effect of Tunnelling Adjacent to Pile Raft Foundations
Supervisor: Prof Harry Poulos

This thesis describes a study of the tunnelling-induced ground deformations and their effects on adjacent pile foundations. Current design practice to predict tunnelling-induced ground movements is generally based on empirical methods that are subjected to some important limitations. For a ground deformation predicted due to tunnelling to be accurate, the prediction methods should account for the effect of a number of parameters, such as tunnel construction method and tunnel driving details, tunnel depth and diameter, initial stress state, and stress-strain behaviour of the soil around tunnel. In this study, the traditional definition of the ground loss parameter is redefines as "equivalent ground loss parameter" with respect to gap g" parameters and incorporated in to analytical solutions to predict the ground movements around tunnel in clays. In addition, design charts are established to estimate the ground loss values due to tunnel excavation based on tunnel depth and diameter, subsurface soil parameters, and geometry of tunnel boring machine. The applicability of these proposed analytical solutions and the design charts to estimate the ground loss values are then checked with some case records, which encompasses a range of ground conditions from very stiff to soft clays. In general, predicted and observed values are in good agreement.

Numerical analyses and a parametric study were performed to examine the effects of tunnelling-induced ground movements on adjacent pile foundations. The analytical solutions proposed have been incorporated into the numerical computation code used in this study. Through the parametric study, it is shown that the influence of tunnelling on pile response depends on a number of factors, including tunnel geometry, ground loss ratio, soil strength, pile diameter, and ratio of pile length to tunnel cover depth. Simple design charts are established for estimating tunnelling-induced pile responses and may be used in practice to assess the behaviour of existing piles adjacent to tunnelling operation. A published case history has been studied in which the measured lateral pile deflections are compared with those estimated from the design charts and fair agreement found.

One of the main thrust of this research program is to obtain experimental data, with a total of three centrifuge model tests undertaken to obtain accurate and detailed data in order to check the applicability of the proposed analytical solutions which predict the tunnelling-induced ground deformations and the proposed analytical solutions which predict the tunnelling-induced ground deformations and the proposed design charts that estimate the tunnelling-induced pile behaviour. The tunnel excavation procedure has been modelled by "equivalent ground loss values" and the specified ground loss values have been achieved by reducing the diameter of the model tunnel. Model piles ( a single pile and a 2x2 pile group) have been driven into the pre-consolidated Kaolin clay sample by a miniature pile driving actuator. The performance of the pile group and the single pile are compared at the same time in the same package. Tunnel depths have been varied in each test in order to assess the influence of tunnel depth on movements, axial pile forces, pile settlements, and lateral pile deformations. The models were instrumented substantially to measure tunnelling-induced vertical and lateral ground movements and the tunnelling-induced behaviour of the pile. A model scale of 1/100 was used with a nominal centrifuge acceleration 100g. The centrifuge test results have been found to be internally consistent, and compare well with the ground movements predicted using proposed analytical solutions and pile behaviour estimated from the proposed design charts.

Recommendations regarding the estimation of tunnelling-induced pile behaviour and limitations of existing methods are outlined. Finally, areas requiring further research, or where additional work may be beneficial, are highlighted.

Kitsana Manivong
Synthesis and Design of Smart Project Management Information Systems for Life Cycle Project Management
Supervisor: Prof Ali Jaafari

This thesis is directed towards the synthesis of a unique project management model, known as Life Cycle Project Management (LCPM), which promotes an holistic and integrative project development and management approach. This thesis focuses on the application of LCPM to capital projects, particularly construction projects. The model is both unique and practical. Unique in the sense that it combines two proven techniques (concurrent engineering and life cycle management) and practical because a comparable model has recently been used in the construction industry.

The LCPM is a natural progression from life cycle cost analysis and life cycle management. Life cycle cost analysis of projects has been around for many years, most noticeably in the building industry. The analysis is used as a basis to compare similar projects. It can be applied throughout the project's life to optimise total costs, including operating and maintenance costs. LCPM requires the processing of life cycle cost and performance data and involves many diverse disciplines, who provide the various project life cycle information. The application of LCPM requires an integrative project management framework.

Concurrent engineering (CE) allows participants to integrate their ideas, knowledge, and data during the design, development, manufacturing, procurement, and marketing phases. This technique has been implemented successfully in the manufacturing industry where it has been shown the approach can deliver products to market in very short time frames.

The LCPM approach embodies the concept of concurrent engineering and integrated teamwork. Under this approach projects are treated as business ventures. Clients and participants can establish aligned goals at the outset. The LCPM model allows the establishment of life cycle goals which can facilitate the on-going management of the project. One of the aims is the continuing and optimal resolution of risks and uncertainties, which provide a basis for value-based decision making. The components of LCPM are:
· A model and infrastructure for integrating all project phases into a single phase using integrated teams/decision approaches, incorporating concurrency concepts and utilising some form of the alliance method of project delivery; and
· A consistent and efficient platform for development of the project from concept to completion and through to facility operation.

LCPM:
· Promotes integration of project information with teams across project life cycle;
· Sets business (strategic) objectives as the basis for on-going decision-making and evaluation;
· Evaluates project variables holistically and integrates the execution of project management functions, using business objectives as the basis for evaluation;
· Facilitates scenario analysis, offering an integrated environment to effectively and interactively apply "what-if" planning;
· Provides a consistent framework for interdisciplinary communication and teamwork;
· Promotes early problem detection; and
· Lays the foundation for the integration of the processes of planning, engineering, documentation, procurement, and construction management throughout the project life cycle.

In order to validate the LCPM model, computer programs: Construction Management Information System (CMIS) and Smart Project Management Information System (SPMIS) have been designed and developed. The programs are Windows-based applications that utilise a database management system. Both systems embody object-oriented programming techniques.

This thesis also contains details of case studies conducted to synthesise and validate the LCPM model. This thesis demonstrates that the LCPM model is indeed feasible and if implemented properly can respond optimally to today's highly dynamic and complex projects.

Jiping Pan,
Cyclic Load Induced Displacements of Foundations on Cemented Soils
Supervisor: Dr David Airey

Abstract to come

Lisa Ruffat
Radar Remote Sensing of Ocean Surface
Supervisor: A/Prof Ian Jones

Abstract to come

Samia Sedra
Drying Creep of Concrete
Supervisor: A/Prof Stuart Reid

The thesis mainly aims to investigate the "Pickett Effect", which has been a matter of prolonged controversy. As yet the contribution of a real (intrinsic) drying creep mechanism in relation to this effect has not been clearly established.

Several extensive series of tests have been carried out within the scope of this research. Test specimens were short concrete prisms. The investigation included specimens tested under bi-eccentric loads, in such a manner as to isolate a real drying creep mechanism (if it exists). Strain measurements were obtained for sealed and drying specimens, as well as for different axial and eccentric loads. Further tests were performed to measure drying shrinkage strains for load-free specimens. Three different sizes of specimens were introduced to evaluate the size effect. The experimental investigation resulted in a wealth of data concerning creep and drying strains. Results did not prove a real drying creep mechanism. Statistical analyses were performed to evaluate the effect of microcracking. Several factors were assessed, including the influence of the load level, loading eccentricities and specimen size. Test measurements were compared with a number of prediction models and codes.

Microscopic observations were carried out using the SEM to investigate the microcracking contribution to the "Pickett Effect" phenomenon. The replica technique was used to image the surface microcracks on both loaded and load-free drying specimens. The investigation proved that the applied vertical compressive loads significantly affect the crack orientation and density. However, for the level of loads investigated, microcracks were observed on all specimens.

A numerical model was developed to simulate the concrete behavior under loading and/or drying conditions. The effects of creep and softening were accounted for in the model. The analysis was based on moisture diffusion models. Numerical results were compared with the experimental data for three different specimen sizes. The drying shrinkage strains have shown a size effect that is not readily modelled using consistent material parameters and simple models of shrinkage-induced cracking and strain softening. Thus, it was concluded that additional factors must be involved. Initial results showed that coupling between diffusion and cracking could have a significant effect on apparent shrinkage strains and may possibly play a significant role in relation to the "Pickett Effect".

Finally, conclusions have been presented and recommendations have been given for further investigation.

Sukit Thepmongkorn
Wind-Induced Coupled and Complex Motions of Tall Buildings
Supervisor: Prof Kenny Kwok

A base hinged assembly (BHA) technique has been developed for wind tunnel aeroelastic model studies of tall Buildings with coupled translational-torsional motion and eccentricity between centre of mass and centre of stiffness. The BHA technique has major advantages in terms of simplicity and ability to change the mass, stiffness, structural damping and geometry of the building model, and also to model coupled translational-torsional motion.

The results of wind tunnel aeroelastic model tests of the Commonwealth Advisory Aeronautical Research Council (CAARC) standard tall building using a two-degree-of-freedom BHA model were found to be similar to those measured by the conventional aerolastic stick model for reduced wind velocities ranging from 2 to 20. The along-wind and cross-wind response characteristics are also consistent with the results of wind tunnel aeroelastic model tests from other published records.

Wind tunnel aeroelastic model tests of the CAARC building using a three-degree-of-freedom BHA model were conducted. The experimental results highlight the significant effects of coupled translational-torsional motion, and eccentricity between centre of mass and centre of stiffness, on both the along-wind and cross-wind acceleration responses for reduced wind velocities ranging from 4 to 20. Coupled translational-torsional motion and eccentricity between centre of mass and centre of stiffness also have significant impacts on the amplitude-dependent effect caused by the vortex resonant process, and the transfer of vibrational energy between the along-wind and cross-wind directions. These resulted in either an increase or decrease of each response component, in particular at reduced wind velocities close to a critical value of 10.

Wind tunnel aeroelastic model tests of the CAARC building using the modified three-degree-of-freedom BHA model, which has the ability to measure the torsional displacement response were conducted. The experimental results highlight the significant effects of coupled translational torsional motion and cross-wind/torsional frequency ratio on the along-wind, cross-wind, and torsional displacement responses for reduced win velocities ranging from 4 to 18, especially at the reduced wind velocities close to a critical value of 10. Cross-wind/torsional frequency ratio for the building model also has a significant effect on the transfer of vibrational energy between the cross-wind and torsional directions.

For the incident wind normal to the narrow face of the CAARC building model, the periodic twisting moment, resulting from non-uniform pressure distributions along the side faces of c building model due to the separation of the wind at upstream corners and the reattachment process of the separated shear layers, was found to be the main excitation mechanism in the torsional direction. The torsional response of the building model was substantially increased at the critical reduced wind velocity for torsional motion, i.e. when c frequency of the periodic twisting moment coincides with the torsional natural frequency the building model. As the mean wind speed increased, the torsional motion was further amplified by the interaction between the approaching wind and the building motion, which resulted in torsional flutter instability.

For the incident wind normal to the wide face of the CAARC building model, the torsional motion of the building model at a cross-wind/torsional frequency ratio close to unity was found to be greatly amplified at reduced wind velocities close to a critical value of 10. This is caused by the effect of resonance between the vortex shedding frequency and the torsional natural frequency of the building model.

Eccentricity between centre of mass and centre of stiffness significantly affects the along-wind, cross-wind, and twisting moment responses of the CAARC building model. The experimental results indicated that with only a 10% eccentricity, a considerable increase in the twisting moment response at the centre of stiffness, and the acceleration response at the corner, of the building model was evident.

Tim Wilkinson
Plastic Behaviour of Cold-Formed Rectangular Hollow Sections
Supervisor: Prof Greg Hancock

The aim of this thesis is to assess the suitability of cold-formed rectangular hollow sections (RHS) for plastic design. The project involved an extensive range of tests on cold-formed Grade C350 and Grade C450 (DuraGal) RHS beams, joints and frames. A large number of finite element analyses was also carried out on models of RHS beams. The conclusion is that cold-formed RHS can be used in plastic design, but stricter element slenderness (b/t) limits and consideration of the connections, are required. Further research, particularly into the effect of axial compression on element slenderness limits, is required before changes to current design rules can be finalised.

Bending tests were performed on cold-formed RHS to examine the web and flange slenderness required to maintain the plastic moment for a large enough rotation suitable for plastic design. The major conclusions of the beam tests were:

Some sections which are classified as Compact or Class 1 by current steel design specifications do not maintain plastic rotations considered sufficient for plastic design.
The current design philosophy, in which flange and web slenderness limits are independent, is inappropriate. An interaction formula is required, and simple formulations are proposed for RHS.

Connection tests were performed on various types of knee joints in RHS, suitable for the column - rafter connection in a portal frame. The connection types investigated were welded stiffened and unstiffened rigid knee connections, bolted plate knee joints, and welded and bolted internal sleeve knee joints, for use in RHS portal frames. The ability of the connections to act as plastic hinges in a portal frame was investigated.

The most important finding of the joint tests was the unexpected fracture of the cold-formed welded connections under opening moment before significant plastic rotations occurred. The use of an internal sleeve moved the plastic hinge in the connection away from the connection centre-line thus eliminating the need for the weld between the RHS, or the RHS and the stiffening plate, to carry the majority of the load. The internal sleeve connections were capable of sustaining the plastic moment for large rotations considered suitable for plastic design.

Tests on pinned-base portal frames were also performed. There were three separate tests, with two different ratios of vertical to horizontal point loads, simulating gravity and horizontal wind loads. Two grades of steel were used for comparison. The aims of the tests were to examine if a plastic collapse mechanism could form in a cold- formed RHS frame, and to investigate if plastic design was suitable for such frames.

In each frame, two regions of highly concentrated curvature were observed before the onset of local buckling, which indicated the formation of plastic hinges and a plastic collapse mechanism. An advanced plastic zone structural analysis which accounted for second order effects, material non-linearity and member imperfections slightly overestimated the strength of the frames. The analysis slightly underestimated the deflections, and hence the magnitude of the second order effects. A second order plastic zone analysis, which did not account for the effects of structural imperfections, provided the best estimates of the strengths of the frames, but also underestimated the deflections.

While cold-formed RHS did not satisfy the material ductility requirements specified for plastic design in some current steel design standards, plastic hinges and plastic collapse mechanisms formed. This suggests that the restriction on plastic design for cold-formed RHS based on insufficient material ductility is unnecessary, provided that the connections are suitable for plastic hinge formation, if required.

A large number of finite element analyses were performed to simulate the bending tests summarised above, and to examine various parameters not studied in the experimental investigation. To simulate the experimental rotation capacity of the RHS beams, a sinusoidally varying longitudinal local imperfection was prescribed. The finite element analysis determined similar trends as observed experimentally, namely that the rotation capacity depended on both the web slenderness and flange slenderness, and that for a given section aspect ratio, the relationship between web slenderness and rotation capacity was non-linear.

The main finding of the finite element study was that the size of the imperfections had an unexpectedly large influence on the rotation capacity. Larger imperfections were required in the more slender sections to simulate the experimental results. There should be further investigation into the effect of varying material properties on rotation capacity.

Kenneth Xu
General Analysis of Pile Foundations and Application to Defective Piles
Supervisor: Prof Harry Poulos

This thesis falls naturally into four main topics to form a broad study of a general analysis of pile foundations and its application to defective piles.

Firstly, a general rigorous 3-D elastic numerical analysis for "global" pile foundations (which include multiple pile groups) is developed using the boundary element method. It considers all 6 load/displacement components for each of the piles with arbitrary dimensions, and also incorporates full coupling effects. Via hierarchical structures, twelve non-zero sub-matrices in a global matrix are derived for the basic influence factors. The analysis is implemented via a computer program, GEPAN.

Secondly, via a number of theoretical-derived expressions for "free field" soil movements/stresses being incorporated into the pile-soil governing equations, a unified 3-D coupled boundary element approach is presented to analyze the response of pile foundations subjected to "passive" loadings. It is shown that the method is capable of giving realistic estimates of passive pile behaviour under a variety of situations, including soil shrink/swelling; soil surface surcharge loading; tunnelling; soil movements arising from driving piles; cavity formation in soil; excavation; and retaining wall construction.

Thirdly, via applying GEPAN, the behaviour of pile foundations containing defects in some or all of the piles is analyzed. An extensive parametric study of the load-deformation behaviour within a single pile, a two-pile group and a pile group containing pile defects is performed under general loadings (axial, lateral, moment and torsional). Model testing of defective piles shows that the general analysis can serve as a "benchmark" for more approximate analyses of defective pile behaviour.

Finally, the further development of non-linear analysis for pile foundations is presented. A comprehensive study on simulating the shearing behaviour of frictional materials is carried out. A set of two unified explicit equations and two basic parameters, describing the distortional and volumetric behaviours, are presented independently. The model developed has potential for future application in the analysis of piles and pile groups, making allowance for real soil behaviour.

Honghua Zhang
Finite Layer Method for Analysis of Piled Raft Foundations
Supervisor: A/Prof John Small

Analysis of the behaviour of piles, pile groups, rafts and piled raft foundations subjected to lateral and vertical loading as well as moments has been of particular interest in geotechnical and offshore engineering. The behaviour of such foundations under vertical loading has been extensively investigated by a great number of researchers in the past. However, the behaviour of raft and piled raft foundations subjected to horizontal loading has not received as much attention.

In this thesis, two different methods for analysis of single piles that are based on finite layer theory are discussed: the flexibility matrix method and the stiffness matrix method. Calculation shows that for single piles, both methods yield exactly the same results.

Computer programs based on an interaction method and a stiffness method were also developed for analysis of pile groups. The stiffness method is suitable for dealing with small pile groups while the interaction method can be used to analyse large pile groups. By using the interaction method, the behaviour of a pile group embedded in a layered soil has been carefully examined.

The present finite layer method is also used for the analysis of a raft foundation resting on a soil layer of finite depth (with a rough soil-raft interface) and subjected to horizontal and vertical loadings as well as moments. This method can be used to analyse buildings constructed on raft foundations, or pavements subjected to shear loading due to traffic braking or acceleration.

Piled raft foundations subjected to horizontal and vertical loading as well as moments are of main interest in this thesis. A powerful program APRAF (Analysis of Piled Raft Foundations) has been developed based on the finite layer theory. Use of this program has shown that the present method has the following advantages over existing methods:

It is an accurate numerical method based on mature theories.
It only requires a small amount of computer memory.
It can deal with loads and moments applied in any direction to a piled raft.
All the loads applied to the piled raft and induced displacements are coupled.
Computing time is relatively small.

Comparisons with existing published experimental results show that although the present method is based on elasticity theory, it can well predict the behaviour of pile groups and piled rafts under working loads.

Kan Ding
The Behaviour of Pipes Constructed from Rib-Stiffened Corrugated Structural Plate under Iron Ore Stockpiles
Supervisor Prof John Carter

The structural behaviour of buried culverts provided with rib stiffeners, under the influence of live loading imposed by iron ore stockpiles, has been examined using non-linear elasto-plastic finite element analysis.

Several case studies of engineering projects have been used to identify important aspects of the structural behaviour of the stiffened pipe and their structural interaction with the surrounding soil.

Analyses of the buried structures commonly use two-dimensional finite element techniques assuming plane strain conditions. The modelling of the buried structures considered here has been carried out the geotechnical finite element package AFENA (Carter and Balaam, 1995), in which the soil has been modelled by two-dimensional solid elements and the corrugated plates with ribs by beam-column members.

Several important measures of the stiffened pipe's performance have been investigated, such as the diametrical strains, hoop thrusts, bending moments, hoop and bending stresses, etc. The finite element predictions of tunnel response have been validated with independent finite element predictions reported by Moore (1997) and the results of a field survey supplied by Ingal Civil Products.

The most important parameters that affect the magnitude and distribution of forces and moments around the pipe, as well as the changes in the diametrical strains of the tunnel during the backfill construction have been investigated. These parameters include soil and rib-stiffener properties.

The usual design procedure for a stiffened pipe is based on the current standard (AS2041/2042-1984). Some suggestions are made regarding ways in which the current procedures for a stiffened pipe could be improved. The current design procedure does not recognise the potential problems associated with increasing structural stiffness. Therefore, a more rigorous design procedure directly relevant to rib-stiffened structures is required.

Kim Khoo
Fuzzy Control of Structural Vibration
(Supervisor Prof Kenny Kwok)

This thesis presents the results of a study of the use of fuzzy controllers to control structural vibration under dynamic environmental loading from earthquake as well as wind forces. It compares the dynamic response of a structure under fuzzy control with the response of a structure controlled using classical instantaneous optimal control. The effectiveness of each control method is investigated with regard to the robustness of the response as well as the cost efficiency of the control effort.

Preliminary investigations of the performance of the optimal and fuzzy controller were performed on a simple single degree of freedom system model of a cantilever. A fuzzy controller was designed when satisfactory optimal control was achieved for the system. The maximum control force applied by the fuzzy controller is equivalent to the maximum force expended during optimal control. The structural properties of the system were then altered and the response of the cantilever under dynamic loading was controlled using the original fuzzy and optimal controllers.

The control of two and three dimensional structural systems then followed. The first is a frame consisting of a 1.8m column connected to a 3m column by a rigid beam connection. In order to investigate the effectiveness of the controllers on more complex structures encountered in a civil engineering environment, the research progressed to buildings modelled as cantilevers with lumped masses possessing translational and rotational degrees of freedom at the nodes representing the floor levels of the building. Optimal and fuzzy controllers were then designed for the control of a sixty storey building. Subsequently a benchmark model of a five storey building was controlled.

Earthquake loading was applied to the two-dimensional frame and the five-storey benchmark model. Wind loading obtained from wind tunnel tests was applied to the sixty storey building model. The computer simulation was carried out using the MATLAB Simulink program. The structures were modelled using the finite element modelling package STRAND6. Fuzzy controllers were designed using MATLAB Fuzzy Logic Toolbox after an optimal control solution to the control blem was formulated. The controllers derived the required control force from the displacement and velocity of the structure. The optimal controller was used to calibrate the control effort applied by the fuzzy controller in order to compare the maximum displacements, velocities and accelerations generated during control.

The results of the research show that the fuzzy controller is able to control the structure to a degree comparable to and under certain circumstances better than the optimal controller. The controlled response of the fuzzy controller improved in comparison to the controlled response of the optimal controller, when the model of the actual structure being controlled deviated from the model that the controllers were designed on. The optimal controller still performs better than the logic controller with small changes in the structural property. The fuzzy controller is able to reduce the displacement response of the structure effectively, however it is less effective in controlling the acceleration responses of the structure. Fuzzy controllers are used in the field of structural control for the reason that they are able to achieve a more robust control performance. The results of this thesis confirm that this is true.

The University of Sydney - Civil Engineering

2000 Higher Degree Theses