1998 Higher Degree Theses
Cameron Chick, The Behaviour and Design of Thin-Walled I-Section Beam-Columns
Anthony Hasham, Flexural-torsional buckling of thin walled I-sections subjected to combined compression and major axis bending
Mohammad Jonaidi, Effects of Differential Settlement on Storage Tank Shells
Dan Lange-Kornbak, Micromechanics of concrete
Martin O'Shea, The Behaviour of Thin-Walled Concrete Filled Steel Tubes
Catherine Rousch, Behaviour of Purlin Sheeting Systems
Subramaniam Santhikumar, Fracture of viscoelastic tension softening materials
Jayanta Sinha, Piled Raft Foundation Systems in Expansive Soils
Lip Teh, Towards 3D Plastic-Zone Advanced Analysis of Steel Frames
Ben Young, The Behaviour and Design of Cold-Formed Channel Columns
Xiaosu Zheng, Computational methods in Geomechanics
Fei Zhou, Development of Tent-Roof Shells
Cameron Chick
The Behaviour and Design of Thin-Walled I-Section Beam-Columns
Supervisor: A/Prof Kim Rasmussen
The objective of this thesis to present the results of an investigation into the behaviour of thin-walled I-sections loaded in combined compression and bending. The basis of this work includes the design fabrication and testing of full-scale thin-walled I-section specimens and the development of a geometrically nonlinear finite element analysis.
The experimental study consisted of tests performed on thin-walled I-sections fabricated from high strength steel plates by welding. Tests were performed in pure compression, pure bending and combined compression and bending. The local and overall geometric imperfections of all specimens were measured prior to testing and these values were used in modelling their behaviour. Residual stresses caused by the manufacturing and fabrication processes were measured in typical cross-sections. Stub column tests were performed on short specimens to determine the axial section capacity while the majority of tests were performed in a horizontal reaction frame between pinned ends. Tests were performed on one type of cross-section over three member lengths. The local buckling loads as well as ultimate loads were recorded for each test. Based on values of column slenderness, short, intermediate and long length specimens were tested to examine three modes of failure. These consist of local buckling, overall buckling and interaction of local and overall buckling.
The primary goal of the theoretical study was the development of a geometrically nonlinear finite element analysis based on the formulation of the 9-noded Lagrangian degenerated shell element.
The analysis is capable of modelling the behaviour of
general shell geometries which are subject to concentrated and distributed
loadings. The boundary conditions consist of three translational and three
rotational restraints at each node which may be specified in the global
directions or relative to axes local to the shell surface.
Comparisons have been performed between the experimental ultimate strengths
and the design ultimate strengths of the Australian, European and American
specifications.
Anthony Hasham
Flexural-torsional buckling of thin walled I-sections subjected to combined
compression and major axis bending
Supervisor: A/Prof Kim Rasmussen
The objective of this thesis is to present the results of an investigation into the behaviour of thin-walled I-sections loaded in combined compression and major axis bending. The basis of this work includes the design, fabrication and testing of full-scale thin-walled I-section specimens. The work also includes the development of an inelastic material and geometric nonlinear finite element model whose results are verified against the test results. Furthermore, the theoretical model is applied to beam-columns of various cross-sections in order to determine the shape of the interaction curve for I-sections subjected to combined compression and major axis bending. Most of the cross-sections are slender and suffer local buckling at ultimate. A new modified design interaction curve is proposed for I-sections subjected to combined compression and major axis bending in order to enhance the current design interaction curve according to the rules of the Australian Standards for Steel Structures, AS 4100 (1990).
The experimental study consisted of four series of tests. Series I had a beam-type cross-section with stocky flanges and was braced at mid-length in order to prevent out-of-plane buckling. The Series I tests provided the section capacity with no reduction in load due to overall instability, hence failure was by a combination of in-plane bending and local buckling of the web plate. Series II, III, and IV, had a column-type cross-section with slender web and flanges. Series II was also braced at mid-length in order to obtain the section capacity and failed by in-plane bending and local buckling of the component plates. The Series III and IV specimens were free to buckle out-of-plane at mid-length, hence failure was by a combination of in-plane bending, out-of-plane flexural-torsional buckling, and local buckling of the component plates.
The beam-column tests were performed on thin-walled I-sections fabricated from high strength steel plates by welding. A horizontal reaction frame consisting of two interlinked actuators was used for the tests. Tests were performed in pure compression, pure bending, and in combined compression and bending. Sufficient tests were performed to obtain the experimental axial force versus bending moment interaction curves.
Detailed measurements of the local and overall geometric imperfections of all beam-column specimens were performed prior to testing. Residual stresses caused by the fabrication process were measured on typical cross-sections. Two fix-ended stub column tests for each cross-section were performed on short specimens to determine the axial section capacity.
The local buckling loads as well as the ultimate loads were recorded for all test specimens and compared with the theoretical predictions of the finite element analysis. The experimentally observed behaviour was compared with the theoretical predictions and the accuracy of the theoretical model to describe the observed test behaviour was assessed. The experimental and theoretical interaction curves were convex for all series.
The primary aim of the theoretical study was to develop a nonlinear inelastic finite element model which was able to accurately determine the strengths and behaviour of thin-walled I-section beam-columns subjected to combined compression and major axis bending. The model was used to analyse other I-sections in addition to those tested and an enhanced interaction curve was proposed.
The experimental and theoretical ultimate strengths are compared with the ultimate design strengths predicted by the Australian Standards for Steel Structures, AS 4100 (1990); the American Institute of Steel Construction Load and Resistance Factor Design Specification for Structural Steel Buildings, the AISC-LRFD Specification (1993); and the European Committee for Standardisation, Eurocode 3 (1992). This comparison aimed mainly to investigate whether the shape of the interaction curves given by codes of practice, which are based on tests of compact or semi-compact cross-sections, applied to slender cross-sections. The ability of the specifications to predict accurately the strength of thin-walled I-section beam-columns subjected to combined compression and major axis bending is evaluated, and the philosophy adopted in the specifications for the design of I-section beam-columns is discussed.
Mohammad Jonaidi
Effects of Differential Settlement on Storage Tank Shells
Supervisor: A/Prof Peter Ansourian
In this thesis, attempts have been made to highlight the influence of the differential (uneven) settlement on the shell of closed and open-top steel storage tanks, with special attention given to understanding of the physical response of the practical cylinders, which have non-uniform shell thickness. Settlement is imposed as a vertical translation at the base, generally but not exclusively in the form of harmonic curves. Both theoretical and experimental investigations are carried out to fulfil this aim.
In theoretical approach, inextensional theory, membrane theory and semi-membrane theory have been applied for linear analysis of the shell. A simple expression has been developed for maximum meridional stress resultant at the bottom of the shell of closed-top tanks, verses maximum settlement. Finite Element Analysis (FEA), regarded as the most accurate solution, was used to gain insight into the behaviour of more realistic cases, to assess the accuracy and limitations of the more limited solutions, and to carry out parametric studies at different levels of taper, aspect ratios and slenderness. Sensitivity of closed-top cylinders to circumferential restraints at the top was figured out and then assessed in the numerical study.
The finite element analysis has also been carried out on the refined mesh models both in linear buckling and nonlinear approaches. In a linear buckling parametric study the buckling of cylindrical panels, as a part of whole cylinder, have been investigated leading to developing an expression for critical stresses and critical settlements of cylinders under edge deformation. In limited geometries, the nonlinear and post-buckling behaviour of tanks under harmonic settlements have been studied with attention given to the effect of imperfection and internal pressures on the shell behaviour.
The experimental program has been conducted on seven precisely machined steel specimens of 600 mm diameter with uniform and stepped wall shells by using a test rig completely designed by researcher. Both open and closed top models were tested under low imposed settlement in a linear approach and large settlement in the non-linear and post-buckling study. Particular attention has been given in the measurement of circumferential stresses developed in the girder of open top models and meridional stresses at the base of closed-top models. The buckling mode of shear form have been experienced in the uniform shells and at the top of stepped shells.
Dan Lange-Kornbak
Micromechanics of concrete
Supervisor: Prof Bhushan Karihaloo
Summary of this thesis is not available.
Martin O'Shea
The Behaviour of Thin-Walled Concrete Filled Steel Tubes
Supervisor: A/Prof Russell Bridge
Thick-walled concrete filled steel tubes have been under extensive experimental investigation for many years due to the inherent advantages of this type of member. Recently due to economic considerations, thin-walled steel tubes, often in conjunction with high strength concrete, have become popular (multi-storey building construction, bridge piers and foundations). However there is very little available research which has been conducted on the behaviour of this type of member. In this thesis an extensive experimental investigation into the behaviour of thin-walled steel tubes filled with medium to very high strength concrete is presented.
The response of the unfilled steel tube has been experimentally investigated. The axial load was applied concentrically and for some specimens at small eccentricities. The restraint provided by an internal medium was experimentally modelled for axially loaded tubes. The complete material properties of the thin-walled steel tubes were measured. This included residual stresses and geometric imperfections.
Confinement of medium to very high strength concrete by thin-walled circular steel tubes has been experimentally investigated. Three different axial compression tests were conducted which included: concentrically loading the steel and concrete simultaneously; concentrically loading only the concrete (steel unbonded and only used as confinement); and eccentrically loading the steel and concrete simultaneously at eccentricities of D/20 and D/10, where D is the external diameter of the steel tube.
The results from the experimental investigations have been compared to estimates from several design methods. After modification, these design methods can be used to conservatively estimate the strength of thin-walled steel tubes filled with medium to very high strength concrete under the examined loading conditions.
Catherine Rousch
Behaviour of Purlin Sheeting Systems
Supervisor: Prof Greg Hancock
Roof and wall systems often consist of cold-formed steel channel or Z-section purlins screw-fastened to high tensile profiled steel sheeting. The objective of this thesis is to investigate the behaviour of these roof and wall systems. For this purpose, both theoretical and experimental studies have been performed.
A non-linear elastic purlin analysis, developed for the specific purpose of determining the lateral deflections of, and stresses in, the unconnected flanges of simply-supported and continuous purlins screw-fastened to sheeting and subject to wind uplift loading, is described. The analysis incorporates a purlin design model, which models the distortional behaviour of screw-fastened purlins under wind uplift loading. The purlin design model includes the torsional restraint provided by the sheeting to the purlin. By including the distortional and local (flange-web) buckling failure stresses in the non-linear analysis, purlin failure loads and failure modes can be predicted.
Two experimental investigations are presented. In the first investigation, purlin-sheeting connection tests were performed in an effort to quantify the magnitude of the torsional restraint provided by sheeting to a screw-fastened purlin. In particular, the effect of the location of the screw-fastener on the magnitude of the restraint was examined. In the second investigation, a series of full-scale vacuum rig tests simulating wind uplift loading on simply-supported screw-fastened channel and Z-section purlins was described. The purpose of the tests was to obtain accurate non-linear response data for unbridged (unbraced) purlins, and for purlins with intermediate bridging.
The failure loads, failure modes, lateral deflections and stresses of the screw-fastened purlins in the vacuum rig tests are compared with those predicted by the non-linear analysis. In addition, the failure loads and failure modes of both simply-supported and continuous screw-fastened purlins determined from vacuum rig tests conducted earlier are compared with those predicted by the non-linear analysis. The accuracy of the non-linear analysis in predicting purlin failure loads and failure modes is examined.
From the failure loads predicted by the non-linear
analysis for the simply-supported and continuous screw-fastened purlins
subject to wind uplift loading, R-factors have been calculated. The
calculated R-factors are compared with the R-factors in the proposed
Australian/New Zealand design standard. The potential of using the
non-linear analysis to determine R-factors in order to minimise testing is
discussed, including calibration of the design model for limit states
design.
Subramaniam
Santhikumar
Fracture of viscoelastic tension softening materials
Supervisor: A/Prof Stuart Reid
Summary of this thesis is not available.
Jayanta Sinha
Piled Raft Foundation Systems in Expansive Soils
Supervisor: Prof Harry Poulos
The objective of this thesis is to analyse the behaviour of piled raft foundation systems in expansive soils incorporating the influence of ground movements. The load distribution patterns between the raft and piles have been analysed, and their impact on total and differential settlements and stresses underneath the raft and along the pile shafts have been critically reviewed. The parameters which influence the behaviour of these foundations including number and location of piles, raft stiffness, magnitude of ground movements and applied loadings have also been considered in the analysis. A method of analysis has been developed to carry out various stages of formulation, and a computer program PRAExS (Piled Raft Analysis in Expansive Soils) has been written using FORTRAN language. This analysis has also been extended to include free standing piled groups, ie piled raft systems with a physical gap between the raft and the soil surface. The advantages and disadvantages of free standing pile groups have been compared with those of piled raft foundation systems when subjected to ground movements. To validate the method of analysis, the results have been compared with actual field and laboratory measurements.
A thorough review of literature has been carried out to investigate the development of knowledge at various stages of research. The shortcomings of the present state of knowledge have been recognised.
To understand the complexity of piled raft behaviour, individual behaviour of rafts and piles have been analysed and discussed. Computer programs RAExS (Raft Analysis in Expansive Soils) and PAExS (Pile Analysis in Expansive Soils) have been developed to carry out these analyses incorporating soil-structure interaction at the soil-raft, soil-pile and pile-raft interfaces. Non-linear conditions have been incorporated by considering yielding of soil at the raft and pile bases, slip at soil-pile interfaces and lift-off between the soil and raft.
Finally, the results have been summarised and conclusions presented. Attention has also been drawn towards the practical implications in relation to foundation performances subjected to ground movements. The scope of further research has also been identified and discussed.
The following papers have been published during this thesis work:
* "Behaviour of Stiffened Raft Foundations" by Sinha, J and Poulos, H.G. (1996), 7th ANZ Conference on Geomechanics 1996, Adelaide, Australia, pp 704 - 709.
* "Piled Raft Foundation Systems in Swelling and Shrinking Soils" by Sinha, J and Poulos, H.G. (1997), XIVth International Conference on Soil Mechanics & Foundation Engineering 1997, Hamburg, Germany, pp 1141 - 1144.
* "Comparison of Some Methods for analysis of Piled Rafts" by Poulos, H.G., Small J., Ta, L.D., Sinha, J & Chen, L (1997), XIVth International Conference on Soil Mechanics & Foundation Engineering 1997, Hamburg, Germany, pp 1119 - 1124.
Lip Hen Teh
Towards 3D Plastic-Zone Advanced Analysis of Steel Frames
Supervisor: Dr Murray Clarke
The latest Australian steel structures standard, AS4100-1990, is the first in the world to formally incorporate the provisions for a certain method of second-order inelastic analysis which can model the in-plane strength and stability of frames to such accuracy that the individual specification checks of member capacity are waived. This type of 2-D inelastic analysis is referred to as advanced analysis. Three-dimensional advanced analysis is not provided for in the standard, nor in any other existing standard around the world, apparently due to the perception that some technical problems in practical 3-D advanced analysis of steel frames are still unsolved.
The objectives of this thesis are to develop a computer program which is capable of practical 3-D advanced analysis of steel frames, and to investigate the economic potential of 3-D advanced analysis in the design of steel space frames. In the course of working towards these objectives, efforts have been undertaken to resolve a number of controversies and inconsistencies existing in the literature on the second-order elastic analysis of 3-D framed structures.
A new definition of conservative moment is proposed as
being the correct representation of the internal bending moments and torque
of a spatial elastic beam. This representation resolves the perennial
controversy concerning the behaviour of conservative internal bending
moments under finite rotations in space. The definition of spatial rotation
which is work-conjugate with the vector of internal moments of a spatial
beam is also identified. This identification is an important undertaking
since energy principles are widely used to formulate the governing
equilibrium equation of a beam element. Neglect of this aspect of a spatial
beam by some researchers has resulted in their failure to detect the
flexural-torsional instability of various framed structures.
Irrespective of their conjugateness with the vector of internal moments, the
suitability of different types of spatial rotation in the formulation of
second-order spatial beam elements is also discussed in detail. It is
pointed out that “commutativity” is a necessary but insufficient
condition for a particular type of spatial rotation to be used in the
formulation of a second-order spatial beam element.
Based on the research results mentioned in the preceding paragraphs, an Updated Lagrangian formulation of a spatial elastic beam element is developed. The inconsistency found in the literature regarding the symmetry of element tangent stiffness matrices derived in the Lagrangian frameworks is explained. A Co-rotational formulation of a spatial elastic beam element is also presented, and the validity of the Wagner effect term is confirmed. It is demonstrated that one Co-rotational cubic element is sufficient to model each member of an elastic steel frame in the large displacement analysis with excellent accuracy.
The numerical results obtained using the Updated Lagrangian and the Co-rotational elastic beam elements are used to select the framework in which the 3-D plastic-zone element is formulated, and to justify some of the approximations that are introduced to the formulation. It is demonstrated through numerical examples that quite excessive plastic-zone beam elements have been used by previous researchers, greatly exacerbating the inefficiency of plastic-zone analysis. In general, only 3 elements are required to accurately model a storey column, and 4 or 5 elements are sufficient for a base column with a rigid support.
A series of advanced analyses and second-order elastic analyses/designs are carried out on two steel space frames composed of circular hollow sections. Various shortcomings of the latter procedure are illustrated and explained. The rotation capacity required of circular hollow sections is also investigated in the case study. The theoretical weight savings that can be achieved through the use of advanced analysis to design the steel space frames are computed for different conditions. The issue concerning the appropriate capacity factor to be used in advanced analysis is also discussed.
A simple yet effective branch-switching algorithm for tracing the secondary equilibrium path of elastic framed structures is presented in an appendix. Another appendix contains a concise review on the behaviour and strength of semi-rigid steel frames.
Ben Young
The Behaviour and Design of Cold-Formed Channel Columns
Supervisor: A/Prof Kim Rasmussen
The behaviour of fixed-ended and pin-ended singly symmetric columns is investigated experimentally and theoretically in this thesis. It is shown that local buckling does not induce overall bending of fixed-ended singly symmetric columns, as it does of pin-ended singly symmetric columns. Consequently, local buckling has a fundamentally different effect on the behaviour of fixed-ended and pin-ended singly symmetric columns. These fundamentally different effects of local buckling lead to inconsistencies in traditional design approaches which account for support conditions solely by using effective lengths. Therefore, improved design recommendations for fixed-ended singly symmetric columns are proposed in this thesis.
An experimental investigation on cold-formed channel columns compressed between fixed and pinned ends is presented in this thesis. A series of tests was performed on plain and lipped channels brake-pressed from high strength structural steel sheets. Four different cross-section geometries were tested over a range of lengths which involved pure local buckling, distortional buckling as well as overall flexural buckling and flexural-torsional buckling. The different effects of local buckling on the behaviour of fixed-ended and pin-ended channels were investigated by comparing strengths, load-shortening and load-deflection curves, as well as longitudinal profiles of buckling deformations. Geometric imperfections were measured on nearly all test specimens before testing as well as the material properties and residual stresses were also measured. The experimental local buckling loads of the fixed-ended tests were determined and shown to be in excellent agreement with the theoretical local buckling loads obtained using an elastic finite strip buckling analysis.
A test rig was developed for the test program, such that detailed deformation profiles of the test specimens were successfully obtained during testing. Fixed-ended bearings were used which allowed measurements of the axial load and the bending moments about two perpendicular axes.
A theory is described for the bifurcation analysis of locally buckled members by deriving the governing equations for the fundamental and bifurcated states. The analysis uses elastic and inelastic geometric nonlinear finite strip buckling analyses to determine the tangent rigidities of the locally buckled section The tangent rigidities are substituted into the equations for overall flexural and flexural-torsional buckling to obtain the bifurcation loads. The effect of yielding is highlighted through a parametric study. The members are subject to pure compression and assumed to be geometrically perfect in the overall sense. It is shown using the fundamental state equations that fixed-ended singly symmetric columns exhibit overall bifurcation behaviour while pin-ended columns do not. The overall bifurcation loads obtained from the analysis are compared with tests of fixed-ended channel columns. The variation of the bifurcation loads with the length is shown to be in good agreement with the tests. The bifurcation loads are shown to be sensitive to the magnitudes of local and overall geometric imperfections.
The test strengths are compared with the design strengths predicted using the Australian/New Zealand, American and European specifications for cold-formed steel structures. It is shown that a fixed-ended channel column can be designed using an effective length of half of the column length and assuming the applied force acts at the centroid of the effective section. This result supports the conclusion drawn from the experimental and theoretical studies that the shift of the effective centroid does not induce overall bending in a fixed-ended channel column.
It is shown that the current design rules lead to very
conservative predictions for pin-ended columns. The shift of the
effective centroid, as predicted by the Australian/New Zealand and American
specifications, is shown to be inaccurate compared to the tests for the
lipped channel with slender flanges. This led to the abnormal result
of an increase in column strength with increasing length.
Xiaosu Zheng
Computational methods in Geomechanics
Supervisor: A/Prof John Carter
Summary of this thesis is not available.
Fei Zhou
Development of Tent-Roof Shells
Supervisor: A/Prof Rob Wheen
Summary of this thesis is not available.