Beam-column (BC) joints play an important role in the seismic performance of moment-resisting reinforced concrete (RC) frame structures. Without adequate joint shear reinforcement, BC joints can be the most vulnerable elements during an earthquake. Several techniques for improving the seismic performance of BC joints have been proposed, but they have been criticised for being labour-intensive and/or susceptible to premature debonding. This research explores the application of the deep embedment technique for strengthening a shear-deficient beam-column joint. Two approaches, experimental and finite element (FE) study were conducted. The experiment contained the tests of seven exterior RC BC joints under constant column axial load and a reverse cyclic load at the beam end. Variables considered during the experiments were the material type and embedded reinforcement ratio. The FE study included the modelling of the tested specimens using ABAQUS and parametric study to asses the effect of column axial load, concrete compressive strength and embedded bar size on joint shear strength. The experimental results showed the strengthened specimens had superior global and local behaviour compared to the control one. In addition, the maximum joint shear strength also changes linearly with the variation of the concrete strength, column axial load and embedded bar size.

Initially, two experimental programmes studying the structural behaviour of stainless steel beam-to-open column joints and beam-to-tubular column joints under static loads are reported in detail. The joint configurations tested include flush and extended end plate connections, top and seat cleat connections and top, seat and web cleat connections. The full moment-rotation characteristics are reported in detail. It is observed that the connections displayed excellent ductility, superior than that of equivalent carbon steel connections, and attained loads much higher than the ones predicted by design standards for carbon steel joints. Nonlinear FE models have been developed and validated against the experimental results. The FE models are shown to accurately replicate the experimentally determined, initial stiffness, ultimate resistance, overall moment-rotation response and observed failure modes. In addition, a comprehensive parametric study is conducted. The design rules for stainless steel connections, which are based on the specifications of EN 1993-1-8 for carbon steel joints, are reviewed and are found to be overly conservative in terms of strength and inaccurate in terms of stiffness thus necessitating the development of novel design guidance in line with the observed structural response. Hence, simplified mechanical models in line with the observed response are developed.

The building construction industry is growing all over the world and considered as a labour-intensive industry. It is associated with significant safety risks and losses resulting from major accidents. These critical safety risks are largely due to lack of awareness, which causes poor performance. Furthermore, in construction management projects, risk assessment tools are still widely employed by adopting two traditional parameters, severity of consequence (SC) and probability of occurrence (PO), to analyse the safety risk level. It is not clear, however, whether this analysis can evaluate the safety risk magnitude appropriately, which necessitates the introduction of another parameter, probability of consequence (PC), to improve the risk evaluation. The fuzzy reasoning technique (FRT) is useful for quantifying and dealing effectively with the lack of certainty related to the domain of building construction projects. PC was incorporated into the model which allows safety risks to be assessed correctly. Furthermore, the modified fuzzy analytical hierarchy process (MFAHP) and fuzzy technique for order preference by similarity to ideal solution (FTOPSIS) methods are integrated into a new construction safety risks model for the evaluation of important safety risks. Four specific case studies are employed to illustrate the applicability and performance of the proposed model.

Structural members of reinforced concrete (R/C) buildings designed according to older, less stringent seismic codes are often vulnerable to shear or flexure-shear failure followed by axial failure. Thus, such substandard R/C structures are susceptible to vertical collapse, which pertains to the exceedance of vertical resistance of columns and connecting beams and can lead to the whole structure – or a substantial part of it – undergoing collapse. The largest database of shear and flexure-shear critical R/C columns cycled well beyond the onset of shear failure and/or up to the onset of axial failure is compiled and empirical relationships are developed for key parameters affecting the response of such members after the initiation of shear failure. A novel shear hysteresis model is proposed employing these relationships, based on experimental observations that deformations after the onset of shear failure tend to concentrate in a specific member region. A computationally efficient finite element model of the member-type is proposed, using the above shear hysteretic model and combining it with displacements arising from flexural and bond-slip deformations to get the full lateral force-lateral displacement response. It accounts for the interaction between flexural and shear deformations inside the potential plastic hinges, the distribution of flexural and shear flexibility along the element, as well as the location and extent of post-peak shear damage, without relying on assumptions about the bending moment distribution and avoiding shortcomings of previous beam-column models pertinent to numerical localisation. Thus, the full-range hysteretic response of substandard R/C elements can be predicted up to the onset of axial failure subsequent to shear failure with or without prior flexural yielding, while simultaneously accounting for potential flexural and anchorage failure modes. The proposed model is implemented in a finite element structural analysis software and its predictive capabilities are verified against quasi-static cyclic and shake-table test results of column and frame specimens. The model is shown to be sufficiently accurate not only in terms of total response, but more crucially in terms of individual deformation components. Overall, it is believed that the accuracy, versatility and simplicity of this model make it a valuable tool in seismic analysis of complex substandard R/C buildings. An experimental investigation of shear and flexure-shear critical R/C elements is carried out with the aim of independently validating the beam-column model. Furthermore, an opportunity is provided to verify the model’s underlying assumptions, which is of paramount importance for the reliability of its analytical predictions. The experiments were designed in such a manner as to investigate the effect of vertical load redistribution from axially failing members on the lateral post-peak response of neighbouring columns.

In 1968 a relatively small gas exposition on the 18th floor of the Ronan Point tower building resulted in the partial collapse of the structure. This event highlighted that progress collapse may occur to structures under an accidental loading event. Other events, including the bombing of the Murrah federal building in 1993 in Oklahoma, have resulted in the common design requirement that a structure be capable of surviving the removal of a load bearing element. This approach, often referred to as the sudden column loss scenario, effectively ignores the cause of the damage and focuses on the structure’s response afterwards. The refinement of the analysis varies, with options to include the nonlinear and dynamic behaviours associated with extreme events, or to use simplified linear and static models with factors included to account for the full behaviour. Previous research into progressive collapse has highlighted that providing ductility in the connections, and avoiding brittle failures, is important in ensuring the structure maintains integrity after a column loss event. However, the majority of this work has been focused on the behaviour of steel and Reinforced Concrete (RC) frame structures. As flat slab construction is a popular method for many structures, due to the flexibility it offers for layouts and its low storey heights, it is an important to consider flat slab behaviour in more detail. Furthermore, slab elements behave differently to frame structures due to the Alternative Load Paths (ALPs) that can develop after a column loss via two-dimensional bending mechanisms. Additionally, punching shear failure is a known issue due to the thin section depths. This work addresses the issue of the response of RC flat slab structures after a sudden column loss. As previous case studies have demonstrated that brittle failures may lead to progressive collapse of such structures, a complete understanding of the response is required. The nonlinear behaviour of a slab structure, due to both material and geometric factors, is investigated to determine the additional capacity available beyond the usual design limits. Additionally, the dynamic factors involved, primarily due to inertial effects, are also considered. To achieve this, experimental and numerical studies were conducted. A series of 1/3 scale models of slab substructures were constructed to replicate column loss events. Two types of tests were conducted, a static push down test with a support removed and a sudden dynamic column removal case. Displacements, strains and support reactions were recorded throughout, along with cracking patterns. For the dynamic tests a high speed camera was used to obtain the deflection response in the short time period after removal and to observe the formation of cracks. Comparisons between the two cases allowed determination of the dynamic effects on the response of the system. The experimental programme was then replicated using a Finite Element (FE) model. The results taken from the experimental case were used to validate the material and modelling assumptions made during the numerical simulations. This validated model was finally used to investigate a wider range of variables and assess the response of typical structural arrangements, with particular focus on the nonlinear and dynamic factors involved after a sudden column loss. The experimental and numeral investigations demonstrated that after the loss of a column, flat slab structures can maintain integrity due to a change in the load paths away from the removal location. Although in some cases a large amount of flexural damage to the concrete and reinforcement occurred, such effects did not lead to complete failure. However, during the experimental programme some punching shear failures occurred, usually at the corner column locations. From the numerical analysis, shear forces of over twice the fully supported condition occurred as a result of removing a column, which may exceed the designed capacity. Comparisons between a static and dynamic analysis provides information into a suitable Dynamic Amplification Factor (DAF) for use with simplified modelling approaches. Based on the range of structures considered, the maximum increase in deflections as a result of a sudden removal was 1.62 times the static case, this is less than the commonly used factor of 2.0. Additionally, this factor reduces as the nonlinearity increases due to further damage, with a smallest DAF calculated at 1.39. This factor can be reduced further if the column is not removed instantaneously. Finally, the material strengthening effect, due to high strain rates, was considered with the conclusion that as such effects only make a limited increase in the capacity of the slab and may be conservatively ignored. In conclusion, RC flat slab structures are capable of resisting progressive collapse after the loss of a column. This is primarily due to their ability to develop ALPs. However, while flexural damage is usually fairly minimal, progressive punching shear failure is a critical design condition as it may result in a complete collapse. Furthermore, the inertial effects involved after a sudden removal can increase the damage sustained, although current design methods may be over conservative.

The arch is usually described in engineering text books as a curved structure, which under the action of vertical loads, exerts an inclined pressure on its supports. It is really intermediary between a curved beam and a curved strut, approaching the former or the latter according as the bending stresses or compressive stresses are correspondingly predominant. The theory of masonry arches has been, and is now in an unsatisfactory state owing to several reasons. Firstly, the materials of construction are generally cheap, and consequently, economy is not considered important. Secondly, up to quite a short time ago, investigators persisted in a type of theory which was admittedly indeterminate. Thirdly, the enormous advance in the production of iron and steel in the Nineteenth Century gave a great impulse to the erection of structures made of these materials, and so, although engineers began to be alive to the importance of both experimental data and theory in structural design, the masonry arch was somewhat neglected. Fourthly, there were so many arches already built, that the dimensions of new arches were almost always based on these existing ones to, the detriment of research. There is, therefore, up to the present practically no useful data on which to base a satisfactory theory. Empirical results are certainly necessary as the conditions under which an arch bears its load are too varied to admit of treatment by pure theory.

The application of high strength steels (HSS) in the construction industry can lead to more economic design and profound sustainability benefits. To facilitate their use in modern practice, most international structural design codes have included HSS within their contents. Due to limited test data at the time of publishing, HSS design provisions are largely based on those for mild steel, with some restrictions, due to HSS’s inferior ductility and strain-hardening characteristics. Hence, further investigation on the applicability of such design specifications to HSS is required. To this end, within the present research work the structural performance of high strength steel structures employing square and rectangular hot-finished hollow sections is rigorously investigated. Meticulously generated finite element models of individual structural components are validated against test data and subsequently used for the generation of additional structural performance data through the execution of parametric studies. Implementing the aforementioned methodology, focus is also placed upon the structural performance of HSS trusses, whilst the possibility of applying prestress to them to enhance their behaviour is examined. Based on the obtained results, the suitability of current codified design methods to HSS is assessed and appropriate design recommendations are made.

Energy consumption and Green House Gas (GHG) emissions from the UK built environment are reflective of the wider situation across Europe, where according to the Energy Performance in Buildings Directive (EPBD) "buildings account for 40% of total energy consumption in the Union" (European Commission, 2010). In December 2006 the UK Government announced a rapid transition to 'zero carbon' new buildings, as a key step forward in reducing GHG emissions from the domestic and non-domestic sectors (DCLG, 2006a; Weaver, 2007). The Passivhaus standard is the fastest growing energy performance standard in the world and in a growing number of regions across Europe it has been implemented as a mandatory minimum standard for all new buildings (iPHA, 2013). This thesis investigates the applicability of this low energy standard to the UK context, in comparison to conventional alternatives, by examining four inter-related themes: (i) in relation to climate change policy and the UK Government's plan for all new homes to be zero carbon from 2016; (ii) by addressing the limitations of the climate data currently used to design Passivhaus buildings, and developing a new methodology for creating higher resolution probabilistic climate data; (iii) by exploring the uncertainty about the future performance of Passivhaus dwellings in relation to future overheating risk and thereby proposing methods to improve whole life design optimization; (iv) by investigating the hygrothermal implications for new build and retrofit Passivhaus projects and highlighting areas where current risk assessment methods are inadequate. This thesis has argued that the transfer of the Passivhaus standard, or any advanced energy performance standard, from one country or region to another should be accompanied by an extensive programme of context specific research and application testing. The findings of this research have shown that the implementation of the Passivhaus standard, in its present format, in the UK is not without risk and uncertainty. This thesis concludes that that the majority of such risks can be substantially mitigated, through the incorporation of high resolution probabilistic climatic data, transient hygrothermal assessments and global sensitivity analysis techniques. The energy saving and thermal comfort potential of the Passivhaus approach have been shown to be substantial and therefore merit the challenges involved in addressing its successful implementation.

This work explores the influence of evolving anisotropy on the stress-strain behaviour of unsaturated soils and proposes a new constitutive elasto-plastic model for unsaturated soils accounting for evolving anisotropy. An extensive campaign of laboratory tests on both isotropically and anisotropically compacted soil samples under a wide range of stress paths was performed. These experimental data were then employed in developing the new model and investigating its performance. A programme of controlled suction triaxial testing was performed on unsaturated and saturated samples of Speswhite kaolin prepared by two different methods of compaction: isotropic and anisotropic. Tests involved probing stress paths, to investigate the initial forms of the yield surface for isotropically compacted and anisotropically compacted samples at different suction values, and how the yield surface was altered by plastic straining caused by loading stages or by wetting stages with significant collapse-compression. Tests also included shearing to failure, to investigate critical state conditions. Experimental results were interpreted in terms of mean net stress p ̅, deviator stress q and suction s as stress state variables and, alternatively, interpreted in terms of mean Bishop’s stress (defined as p^*=p ̅+ S_r s), deviator stress q and modified suction (defined as s^*=ns, where n is the porosity). The experimental results showed that fabric anisotropy can evolve during plastic straining even for a soil that starts isotropic but is then loaded to anisotropic stress states. Also, the results showed that fabric anisotropy can evolve during wetting stages that involve collapse-compression. Furthermore, the results showed no apparent influence of initial or evolving anisotropy on the critical state, where both the initially isotropic and initially anisotropic samples, loaded at various stress path slopes, showed nearly the same critical states. Critical states can be represented in the q:p ̅ plane by a series of parallel lines for different values of suction and the constant suction cross-sections of the yield surface can be represented by distorted ellipses in the q:p ̅ plane, intersecting the negative axis at the point of intersection of the corresponding critical state line. Alternatively, critical states can be represented in the q:p^* plane by a single straight line (for all values of suction) passing through the origin, and constant suction cross-sections of the yield surface can be represented in the q:p^* plane by distorted ellipses passing through the origin (suggesting that the yield surface expression is simpler when expressed in terms of q,p^* and s^* rather than in terms of q,p ̅ and s). A new constitutive model was formulated in terms of Bishop’s stresses and modified suction based on the above observations and other considerations such as that representing the coupling between mechanical and water retention behaviour is easier with Bishop’s stress than with net stress. The new anisotropic model combines features from the isotropic model for unsaturated soils of Wheeler et al. (2003a) with features for modelling of anisotropy taken from the anisotropic model for saturated soils S-CLAY1. The new anisotropic constitutive model was developed solely as a mechanical model, unlike the constitutive model of Wheeler et al. (2003a), which is a combined mechanical and water retention model. Model simulations of mechanical behaviour with the new anisotropic model were performed by using experimental values of S_r (with no attempt to predict values of S_r), because it was then possible to check whether mechanical aspects of the model were performing well. Model simulations showed that significant improvement in the accuracy of the predicted soil behaviour was achieved by incorporating the role of evolving fabric anisotropy. However, model performance appears more satisfactory in simulating soil behaviour under unsaturated conditions than under saturated conditions. Also, the model is not entirely successful in predicting some aspects of anisotropic soil behaviour such as differences in initial specific volume between isotropically and anisotropically compacted samples.

This work is concerned with the thermal and structural behaviour of reinforced concrete members in fire conditions. The numerical analyses of temperature histories and mechanical behaviour of reinforced concrete structural members subjected to fire are the major components of this research. In this thesis a non-linear finite element procedure is proposed to predict the temperature distribution history in the cross section of structural members, such as beams composed of reinforced concrete, in fire conditions. A theoretical analysis of heat and moisture transfer in concrete was made which incorporated the simplifications that energy transfer by convection and diffusion in concrete could be neglected. However, the effect of water evaporation in concrete was considered. The thermal properties of concrete were considered as temperature and moisture dependent and the thermal properties of steel as temperature dependent only. The fire conditions were described by standard time-temperature fire curves and convection and radiation boundary conditions were used. In order to validate the model a series of verification tests have been carried out through a quantitative comparison of the model predictions against known test results. Fairly good accuracy has been found. A non-linear finite element procedure for predicting the structural behaviour of the planar reinforced concrete members is also developed. The proposed procedure is based on "plane stress" theory and an iterative, secant stiffness formulation is employed. The complex features of structural behaviour in fire conditions, such as thermal expansion, shrinkage, creep, transient strains, cracking or crushing and change of material properties with temperature are considered in this model. Predictions from the model proposed are compared against experimental results, as well as against the model proposedb y previous researchers, and a better correlation to experimental data is found. It is shown that the secant stiffness approach can provide good numerical stability for the analysis of planar reinforced concrete members in fire conditions. The model proposed in this study has the potential to predict the fire resistance of a planar reinforced concrete members with an accuracy that is adequate for practical purposes if realistic material properties are available.