This thesis focuses on closed-cell aluminium foams (ALPORAS and ALULIGHT) and on sandwich panels comprising these foams laminated with 2/2 twill carbon fibre (MTM56/0300) skins. The thesis experimentally and analytically investigates the response of foam-only panels (ALPORAS) to indentation with various indenter sizes and shapes; and also studies the behaviour of sandwich panels to contact damage caused quasi-statically or by impact. Quasi??static uniaxial compression testing is used to determine the mechanical properties of the foams (ALPORAS and ALULIGHT). It is revealed that the plastic collapse strength (σ* pl) obtained from the stress??strain curves is lower than the values predicted by the Gibson-Ashby theoretical model. This phenomenon is explained by the fact that the aluminium foams tested are imperfect, non-homogeneous and non-isotropic, and show a distinct cell elongation. Whereas, the Gibson-Ashby theoretical model was based on the finite element method applied to the response of a unit tetrakaidecahedral closed cell having flat faces. The experimental work shows that the deformation of the foam-only panels to indentation is caused by progressive crushing of the cell bands and by shearing and tearing of the cell walls. This thesis presents new analytical models for the response of the foam-only panels and estimates the applied deformation load in all types of indentation. By fitting the experimental load-displacement curves, the shear strength (τ* pl) and the tear energy (γ) are deduced. Compared to the literature, more consistent results are obtained for the shear strength (τ * pl) and the tear energy (γ) from all types of indentation. It is also suggested to determine (τ * pl) and (γ) through indentations with long punches (FEP and LCP), instead of hemi-spherical or cylindrical indenters, because indentation on enclosed areas shows some indenter size dependence. It is concluded that thinner panels are not suitable for the determination of the tear energy (γ) since the densification of the foam is achieved before the tear resistance is fully engaged. Another objective of this thesis is to study the response of sandwich panels comprising a closed??cell aluminium foam core and laminated with carbon fibre skin to quasi-static and impact local damage. Special attention is paid to the residual (remnant) strength in bending of the already indented sandwich panels (quasi-statically or by impact) up to the failure point. The remnant strength in bending is determined by carrying out four point bending strength tests. The local damage is located on either the compressive or on the tensile side of the sandwich panels. Thus, the capacity of the panels to resist transverse loads after they have been locally damaged at contact is investigated. The contact damage on the sandwich panels is experimentally simulated using spherical indenters. The quasi-static indentation is carried out at a low constant velocity (0.5mm/min) ?? the induced contact damage is found to be independent on the sample thickness but dependent on the indenter diameter. On the contrary, the impact test indicates velocity-dependence of the failure mode of the sandwich panel (i.e. skin breakage or punch through) which is found from the load-displacement curves. The results reveal that there is a correlation between the area of the contact damage and the remnant strength, and that the use of metal foam cores leads to high contact damage resilience of composite structures.

This thesis focuses on closed-cell aluminium foams (ALPORAS and ALULIGHT) and on sandwich panels comprising these foams laminated with 2/2 twill carbon fibre (MTM56/0300) skins. The thesis experimentally and analytically investigates the response of foam-only panels (ALPORAS) to indentation with various indenter sizes and shapes; and also studies the behaviour of sandwich panels to contact damage caused quasi-statically or by impact. Quasi??static uniaxial compression testing is used to determine the mechanical properties of the foams (ALPORAS and ALULIGHT). It is revealed that the plastic collapse strength (σ* pl) obtained from the stress??strain curves is lower than the values predicted by the Gibson-Ashby theoretical model. This phenomenon is explained by the fact that the aluminium foams tested are imperfect, non-homogeneous and non-isotropic, and show a distinct cell elongation. Whereas, the Gibson-Ashby theoretical model was based on the finite element method applied to the response of a unit tetrakaidecahedral closed cell having flat faces. The experimental work shows that the deformation of the foam-only panels to indentation is caused by progressive crushing of the cell bands and by shearing and tearing of the cell walls. This thesis presents new analytical models for the response of the foam-only panels and estimates the applied deformation load in all types of indentation. By fitting the experimental load-displacement curves, the shear strength (τ* pl) and the tear energy (γ) are deduced. Compared to the literature, more consistent results are obtained for the shear strength (τ * pl) and the tear energy (γ) from all types of indentation. It is also suggested to determine (τ * pl) and (γ) through indentations with long punches (FEP and LCP), instead of hemi-spherical or cylindrical indenters, because indentation on enclosed areas shows some indenter size dependence. It is concluded that thinner panels are not suitable for the determination of the tear energy (γ) since the densification of the foam is achieved before the tear resistance is fully engaged. Another objective of this thesis is to study the response of sandwich panels comprising a closed??cell aluminium foam core and laminated with carbon fibre skin to quasi-static and impact local damage. Special attention is paid to the residual (remnant) strength in bending of the already indented sandwich panels (quasi-statically or by impact) up to the failure point. The remnant strength in bending is determined by carrying out four point bending strength tests. The local damage is located on either the compressive or on the tensile side of the sandwich panels. Thus, the capacity of the panels to resist transverse loads after they have been locally damaged at contact is investigated. The contact damage on the sandwich panels is experimentally simulated using spherical indenters. The quasi-static indentation is carried out at a low constant velocity (0.5mm/min) ?? the induced contact damage is found to be independent on the sample thickness but dependent on the indenter diameter. On the contrary, the impact test indicates velocity-dependence of the failure mode of the sandwich panel (i.e. skin breakage or punch through) which is found from the load-displacement curves. The results reveal that there is a correlation between the area of the contact damage and the remnant strength, and that the use of metal foam cores leads to high contact damage resilience of composite structures.

Cellular materials like aluminum foam which is the subject of interest here are generally characterized by high energy absorption capacity per unit weight. Materials of this category can be ideal for applications such as packaging and vehicle body structures for enhanced impact safety. A particularly well-known variety of closed-cell aluminum foam is designated as Alporas, which is studied here. From a viewpoint of mechanical behavior, the foam being considered can be represented using either a detailed cellular approach capturing the voids present in foam structure or a phenomenological approach in which experimental stress-strain response is assigned a-priori to solid elements filling up the space occupied by a foam geometry. Both modeling approaches are studied in the present work. It has been shown for the first time that stress-strain behavior under compression including densification can be predicted well with a Kelvin cell-based model, although scope for further improvement exists. Based on a novel combination of compression tests at low strain rates in a UTM and medium strain rates in low velocity impact tests, a relation between foam strength and strain rate has been proposed. This effect of strain rate on strength is captured in a finite element model for analysis using an explicit code with contact simulation capabilities and the predictions for projectile impact tests at higher strain rates using a gas gun-based device have been found to match commendably with results obtained from the said tests.

Aluminium foam has a range of properties that are desirable in many applications. These properties include good stiffness and strength to weight ratios, impact energy absorption, sound damping, thermal insulation and non combustibility. Many of these characteristics are particularly attractive for core materials within sandwich structures. The combination of aluminium foam cores with thermoplastic composite skins is easily manufactured and has good potential as a multifunctional sandwich structure useful in a range of applications. This thesis has investigated the flexural behaviour of such structures using a combination of experimental and modelling techniques. The development of these structures towards commercial use requires a thorough understanding of the deformation and strain mechanisms of the structure, and this will, in turn, allow predictions of their structural behaviour in a variety of loading conditions. ¶ The experimental research involved the use of an advanced 3D optical measuring technique that produces realtime, full-field strain evolution during loading. This experimental characterisation of strain evolution in this class of sandwich structure under flexural loading is the first of its kind in the world. The experimental work studied the sandwich structure undergoing four-point bend testing. Initial studies compared the behaviour of the aluminium foam structure with a more traditional polymer foam sandwich structure. The aluminium foam structure was found to have equivalent or improved mechanical properties including more ductile deformation and an enhanced energy absorption. An investigation was conducted on the effect of core and skin thickness on the metal structure and a range of flexural behaviours were observed. Analysis of the strain distribution showed a complex development including localised effects from the non-uniform cellular structure of the material. An understanding of the variation with size is important in establishing design methods for utilising these structures. In particular, it is desirable that finite element simulations can be used to predict behaviour of these structures in a diverse range of loading conditions. This aspect was considered in the second half of this study. An existing constitutive model for aluminium foam, developed for use in compression energy absorption studies, was used to investigate finite element simulations of the flexural behaviour of the sandwich structure. The FE model was able to predict the general deformation behaviour of the thinner skinned structures although the magnitude of the load-displacement response was underestimated. It is suggested this may be related to the size effect on the input parameter characterisation. The strain distribution corresponded well with the experimental strain measurements. It was found a simple increase in the material model input parameters was able to more closely match the magnitude of the load-displacement response while maintaining the appropriate strain distribution. The general deformation shape of the model with the thicker skin corresponded reasonably well with the experimental observations. However, further work is necessary on the element failure criterion to capture the shear cracking observed. The strain distributions of the model predicted this failure with high strain concentrations matching those of the experimental contours. The last part of the thesis describes a parametric study on the effect of the foam material model input parameters on the flexural behaviour of the sandwich structure model. An important conclusion of this work is that this material model for aluminium foam can, with some development, be utilized to provide a viable method for simulating aluminium foam composite sandwich structures in flexural loading situations.

ABSTRACT A study is carried out on the production of aluminum foams via powder processing. The study deals mainly with the effect of TiH2 particle size on the process of foaming. Mainly two TiH2 particle sizes were used / namely 27,5 &amp / #61549 / m and 8,5 &amp / #61549 / m. Foaming experiments were carried out at temperatures between 675oC &ndash / 840oC. The viscosity of the system is adjusted by controlled addition of Al2O3. The study shows that choice of foaming agent size is influential in the foaming process. With the use of fine foaming agent, temperatures in excess of 800oC would be required for successful foaming. The study further showed that the relation between foaming and viscosity was also dependent on the particle size. Viscosity of 2.3 mPa.s was found to be a limiting value for successful foaming with fine foaming agent. This value appears to increase with increasing particle size. An analysis is presented with regard to temperature dependence of foaming which takes into account the effect of particle size.

Metal foams are lightweight structures and have large use in many components acting as impact energy absorbers. They have exceptional mechanical, thermal and acoustic properties. The design or selection of foam for packaging is done on the basis of impact loads to be sustained or energy to be absorbed. For transportation of nuclear material, metal foams can be used as a packaging material. It may be noted that apart from other qualification requirements, a package containing nuclear material, has to be certified for drop test. Foam can serve the purpose by providing proper cushioning. Metal foams are still under development and need to be accurately characterized in terms of their mechanical properties as well as cell morphology. The aim of this work is to develop, characterize and model foam using experiments and analysis. Aluminum foam has been developed by powder metallurgy technique and the effect of addition of varying amounts of Mg and Alumina on the strength and energy absorption has been studied. Foams of varying densities have also been developed. The reason for going for higher density is to obtain higher plateau stress. If a package is designed with lower density foam, it may become very bulky and even impractical. The characterization part of the work includes study of porosity distribution, cell wall structure, microscopy, SEM images, etc. Mechanical testing (uniaxial compression) was performed on foam samples to get load deflection curve of foams. Area under a given curve i.e. energy absorbed per unit volume has been compared for various compositions and densities. The analysis part of the work presents effect of specimen size on bulk properties of foam. 2D honeycomb and 3D cases have been discussed. To model the porosities, spherical cavities have been assumed. Uniaxial compression cases with different combinations of porosities have been analyzed. The properties like Young’s modulus, plateau stress, Poisson’s ratio, tangent modulus, etc. have been evaluated. The effect of variation in yield strength and tangent modulus of base material on foam has been studied. It appears that if the model is based on uniform porosity distribution, it may lead to lower bound values of physical properties and give conservative result. Although some of these trends have been observed in published literature, the current numerical study has generated additional information and insight.

Pure aluminium and aluminium-5wt % TiO2 aluminium foams were produced by powder metallurgy technique with the use of TiH2 as foaming agent. Two sizes of TiH2 were used: 20&micro / m and 3&micro / m. It has been confirmed that high level of compaction is the primary requirement in foaming. It was shown that hot swaging could be used as a method of compaction for foaming as it leads to values close to full density. Pure aluminium foamed at 675&deg / C and 725&deg / C leads to a volume expansion between 90-180 %. A model was developed for pure aluminium to explain the pore initiation and the resultant pore size. The model predicts a critical particle size for TiH2 below which bubbles could not form. The size appears to be in the neighborhood of 30&micro / m for 675&deg / C and 6&micro / m for 725&deg / C and is temperature dependent. Equilibrium pore size appears to be a function of TiH2 particle size and not affected significantly by the temperature of foaming. It has also been shown that depth effect, i.e. hydrostatic pressure of liquid metal, is unimportant in foaming process and can be neglected. According to the model, to produce pores of fine sizes, two requirements must be met: use of fine foaming agent and the use of high foaming temperature. Al-5 wt % TiO2 was foamed at 750&deg / C and 800&deg / C, i.e. at temperatures that yield viscosities similar to pure aluminium. The structure of foamed metal and level of foaming, 120-160%, was similar to pure aluminium. Unlike pure aluminium, internal reactions are dominant feature of TiO2 stabilized systems. Solid content of the system increases as a result of internal reactions between Al-Ti and Al- TiO2. When this change occurs, however, is not known. It is possible that the viscosity of the system may be four times of its original value.

Terrorism has become a serious threat in the world, with bomb attacks carried out both inside and outside buildings. There are already many unreinforced masonry buildings in existence, and some of them are historical buildings. However, they do not perform well under blast loading. Aiming on protecting masonry buildings, retrofitting techniques were developed. Some experimental work on studying the effect of retrofitted URM walls has been done in recent years; however, these tests usually cost a significant amount of time and funds. Because of this, numerical simulation has become a good alternative, and can be used to study the behaviour of masonry structures, and predict the outcomes of experimental tests. This project was carried out to find efficient retrofitting technique under blast loading by developing numerical material models. It was based on experimental research of strengthening URM walls by using retrofitting technologies under out-of-plane loading at the University of Adelaide. The numerical models can be applied to study large-scaled structures under static loading, and the research work is then extended to the field of blast loading. Aiming on deriving efficient material models, homogenization technology was introduced to this research. Fifty cases of numerical analysis on masonry basic cell were conducted to derive equivalent orthotropic material properties. To study the increasing capability in strength and ductility of retrofitted URM walls, pull-tests were simulated using interface element model to investigate the bond-slip relationship of FRP plates bonded to masonry blocks. The interface element model was then used to simulate performance of retrofitted URM walls under static loads. The accuracy of the numerical results was verified by comparing with the experimental results from previous tests at the University of Adelaide by Griffith et al. (2007) on unreinforced masonry walls and by Yang (2007) on FRP retrofitted masonry walls. To study the de-bonding behaviours of retrofits bonded to masonry, and find appropriate solution to protect certain masonry walls against blast loading, various retrofitting technologies were examined. The simulation covers explosive impacts of a wide range of impulses. Based on this work, pressure-impulse diagrams for different types of retrofitted URM walls were developed as a design guideline for estimating the blast effect on retrofitted masonry walls. The outcomes of this research will contribute to the development of numerical simulation on modelling retrofitted URM walls, improving the technique for explosion-resistant of masonry buildings, and providing a type of guideline for blast-resistant design. / http://proxy.library.adelaide.edu.au/login?url= http://library.adelaide.edu.au/cgi-bin/Pwebrecon.cgi?BBID=1349719 / Thesis (M.Eng.Sc.) - University of Adelaide, School of Civil, Environmental and Mining Engineering, 2009

Terrorism has become a serious threat in the world, with bomb attacks carried out both inside and outside buildings. There are already many unreinforced masonry buildings in existence, and some of them are historical buildings. However, they do not perform well under blast loading. Aiming on protecting masonry buildings, retrofitting techniques were developed. Some experimental work on studying the effect of retrofitted URM walls has been done in recent years; however, these tests usually cost a significant amount of time and funds. Because of this, numerical simulation has become a good alternative, and can be used to study the behaviour of masonry structures, and predict the outcomes of experimental tests. This project was carried out to find efficient retrofitting technique under blast loading by developing numerical material models. It was based on experimental research of strengthening URM walls by using retrofitting technologies under out-of-plane loading at the University of Adelaide. The numerical models can be applied to study large-scaled structures under static loading, and the research work is then extended to the field of blast loading. Aiming on deriving efficient material models, homogenization technology was introduced to this research. Fifty cases of numerical analysis on masonry basic cell were conducted to derive equivalent orthotropic material properties. To study the increasing capability in strength and ductility of retrofitted URM walls, pull-tests were simulated using interface element model to investigate the bond-slip relationship of FRP plates bonded to masonry blocks. The interface element model was then used to simulate performance of retrofitted URM walls under static loads. The accuracy of the numerical results was verified by comparing with the experimental results from previous tests at the University of Adelaide by Griffith et al. (2007) on unreinforced masonry walls and by Yang (2007) on FRP retrofitted masonry walls. To study the de-bonding behaviours of retrofits bonded to masonry, and find appropriate solution to protect certain masonry walls against blast loading, various retrofitting technologies were examined. The simulation covers explosive impacts of a wide range of impulses. Based on this work, pressure-impulse diagrams for different types of retrofitted URM walls were developed as a design guideline for estimating the blast effect on retrofitted masonry walls. The outcomes of this research will contribute to the development of numerical simulation on modelling retrofitted URM walls, improving the technique for explosion-resistant of masonry buildings, and providing a type of guideline for blast-resistant design. / http://proxy.library.adelaide.edu.au/login?url= http://library.adelaide.edu.au/cgi-bin/Pwebrecon.cgi?BBID=1349719 / Thesis (M.Eng.Sc.) - University of Adelaide, School of Civil, Environmental and Mining Engineering, 2009

Terrorism has become a serious threat in the world, with bomb attacks carried out both inside and outside buildings. There are already many unreinforced masonry buildings in existence, and some of them are historical buildings. However, they do not perform well under blast loading. Aiming on protecting masonry buildings, retrofitting techniques were developed. Some experimental work on studying the effect of retrofitted URM walls has been done in recent years; however, these tests usually cost a significant amount of time and funds. Because of this, numerical simulation has become a good alternative, and can be used to study the behaviour of masonry structures, and predict the outcomes of experimental tests. This project was carried out to find efficient retrofitting technique under blast loading by developing numerical material models. It was based on experimental research of strengthening URM walls by using retrofitting technologies under out-of-plane loading at the University of Adelaide. The numerical models can be applied to study large-scaled structures under static loading, and the research work is then extended to the field of blast loading. Aiming on deriving efficient material models, homogenization technology was introduced to this research. Fifty cases of numerical analysis on masonry basic cell were conducted to derive equivalent orthotropic material properties. To study the increasing capability in strength and ductility of retrofitted URM walls, pull-tests were simulated using interface element model to investigate the bond-slip relationship of FRP plates bonded to masonry blocks. The interface element model was then used to simulate performance of retrofitted URM walls under static loads. The accuracy of the numerical results was verified by comparing with the experimental results from previous tests at the University of Adelaide by Griffith et al. (2007) on unreinforced masonry walls and by Yang (2007) on FRP retrofitted masonry walls. To study the de-bonding behaviours of retrofits bonded to masonry, and find appropriate solution to protect certain masonry walls against blast loading, various retrofitting technologies were examined. The simulation covers explosive impacts of a wide range of impulses. Based on this work, pressure-impulse diagrams for different types of retrofitted URM walls were developed as a design guideline for estimating the blast effect on retrofitted masonry walls. The outcomes of this research will contribute to the development of numerical simulation on modelling retrofitted URM walls, improving the technique for explosion-resistant of masonry buildings, and providing a type of guideline for blast-resistant design. / http://proxy.library.adelaide.edu.au/login?url= http://library.adelaide.edu.au/cgi-bin/Pwebrecon.cgi?BBID=1349719 / Thesis (M.Eng.Sc.) - University of Adelaide, School of Civil, Environmental and Mining Engineering, 2009