NUMERICAL CHARACTERISATION OF HOLLOW SPHERE COMPOSITES BASED ON PERFORATED INCLUSIONS MOHD AYUB BIN SULONG UNIVERSITI TEKNOLOGI MALAYSIA NUMERICAL CHARACTERISATION OF HOLLOW SPHERE COMPOSITES BASED ON PERFORATED INCLUSIONS MOHD AYUB BIN SULONG A thesis submitted in fulfilment of the requirements for the award of the degree of Master of Engineering (Mechanical) Faculty of Mechanical Engineering Universiti Teknologi Malaysia NOVEMBER 2011 iii The credit of this work goes to my eminent supervisor, PROFESSOR DR.-ING. ANDREAS ÖCHSNER D.SC. for his generosity in guiding me through the research, to my beloved parents and wife for their motivational support. iv ACKNOWLEDGEMENT I would like to thanks all of my friends who have supported me during completing this work, especially to Meysam, Moones and Huda which helped me to get everything started. I also would like to express my gratitude to UTM and FKM members who have supported continuously throughout this research, Research Management Center (RMC) for the facilities provided, Office of Postgraduate Studies (SPS) for their great administration task, UTM library for striving to serve the best service of resources. Last but not least many thanks to Malaysian Ministry of Higher Education (MOHE) for supporting this research financially. v ABSTRACT Metallic hollow sphere structures (MHSS) are a new type of reinforced materials and can be classified as an advanced composite material. A modified metallic hollow sphere MHS geometry which introduced the perforation becomes the main model in this research. This structure is called a perforated hollow sphere structures (PHSS) which is opened to be infiltrated by the matrix to fully embed it and form a composite. PHSS composites offer a new field of mechanical properties compared to cellular structures studied by other researchers. Emphasis will be given to determine the influence of the modified perforation diameter of PHSS composite in terms of macroscopic mechanical properties (e.g. Young’s modulus and Poisson’s ratio). In addition, the mechanical properties of PHSS composites were also compared to hollow sphere (HS) composites (with and without filled matrix). A perforation introduced in the sphere shells obviously changes the mechanical properties of the PHSS composite, e.g. Young’s modulus and Poisson’s ratio. The result of the investigation revealed that these values decrease as the perforation diameter increases. PHSS composite models were simulated based on the unit cell approach by means of the Finite Element (FE) method. This method can reduce the costs of experimental tests and provides more information on possible mechanical properties of perforated hollow sphere structures (PHSS) composites. Nevertheless, experimental tests are still necessary and should be conducted in the future for validation purpose. vi ABSTRAK Struktur Sfera Logam Berongga adalah jenis baru bahan pengukuh dan boleh dikelaskan sebagai bahan komposit termaju. Geometri Sfera Logam Berongga yang telah diubahsuai iaitu mempunyai lubang menjadi model utama bagi kajian ini. Struktur ini dipanggil sfera berongga berlubang terbuka untuk dimasuki oleh matriks untuk menerapkan sepenuhnya dan membentuk bahan rencam. Komposit sfera berongga berlubang menawarkan satu sifat baru mekanikal berbanding struktur sel yang dikaji oleh penyelidik lain. Tumpuan kajian ini adalah untuk menentukan pengaruh diameter penembusan komposit sfera berongga berlubang yang telah diubahsuai dari segi ciri-ciri makroskopik mekanikal (contohnya modulus Young dan nisbah Poisson). Di samping itu, sifat-sifat mekanik komposit sfera berongga berlubang juga dibandingkan dengan komposit sfera berongga (dengan atau tanpa matriks isian). Penembusan yang diperkenalkan dalam cengkerang sfera merubah sifat-sifat mekanik komposit sfera berongga berlubang dengan ketara, contohnya Modulus Young dan nisbah Poisson berkurangan kerana kenaikan diameter penembusan. Model komposit sfera berongga berlubang disimulasikan berdasarkan pendekatan sel unit dengan menggunakan analisis kaedah unsur terhingga. Kaedah ini boleh mengurangkan kos ujian ujikaji dan memberikan maklumat lanjut mengenai sifat-sifat mekanikal yang mungkin bagi komposit sfera berongga berlubang. Walau bagaimanapun, ujikaji sebenar masih diperlukan dan perlu dijalankan pada masa hadapan bagi tujuan pengesahan. vii TABLE OF CONTENTS CHAPTER TITLE PAGE DECLARATION ii DEDICATION iii ACKNOWLEDGEMENT iv ABSTRACT v ABSTRAK vi TABLE OF CONTENTS vii LIST OF TABLES x LIST OF FIGURES xi LIST OF ABBREVIATIONS xvii LIST OF SYMBOLS xviii LIST OF APPENDICES 1 xxi INTRODUTION 1 1.1 Introduction 1 1.2 Problem Identification 2 1.3 Objective 3 1.4 Scope of Study 3 viii 2 1.5 Gantt Chart 4 1.6 Summary 5 LITERATURE REVIEW 6 2.1 Introduction 6 2.2 Prospective for Lightweight Cellular Metals 6 2.2.1 High specific stiffness and strength 7 2.2.2 Good Energy Absorbers 9 2.2.3 Good Thermal Conductors 9 2.2.4 Structural Vibration 2.3 Perforated Hollow Sphere Structures 11 2.4 Literature Assessment 15 2.5 Finite Element Method 18 2.5.1 Representative Volume Element 3 10 19 2.6 Summary 20 RESEARCH METHODOLOGY 21 3.1 Introduction 21 3.2 Modelling of PHSS 21 3.3 Finite Element Approach: Geometry, Mesh, 27 Boundary Conditions and Materials 3.3.1 Geometry 28 3.3.2 Mesh 35 3.3.3 Boundary Conditions 38 3.3.4 Materials 42 3.4 Mechanical Properties 44 3.4.1 Young’s Modulus and Poisson’s Ratio 44 3.4.2 Deformation of the Model 46 3.5 Thermal Properties 47 3.6 The average Density for Composite materials 49 3.7 Summary 51 ix 4 5 RESULTS & DISCUSSION 52 4.1 Introduction 52 4.2 Young’s Modulus and Poisson’s Ratio 52 4.3 Thermal Conductivity 60 4.4 Summary 65 CONCLUSIONS AND FUTURE STUDY 66 5.1 Conclusions 66 5.2 Future Study 67 REFERENCES 71 APPENDICES A-B 78 x LIST OF TABLES TABLE NO. TITLE PAGE 3.1 Summary of the considered geometry................................. 30 3.2 Specific mechanical properties of the spheres and binder.. 42 3.3 Heat conductivity properties of base materials................... 42 xi LIST OF FIGURES FIGURE NO. 1.1 TITLE PAGE Cellular metals: a) M-Pore® (aluminium sponge);.............. 2 b) Alporas® (aluminium foam); c) Brass foam 1.2 Gantt chart..................................................................... 4 2.1 Properties of cellular metals............................................. 7 2.2 Evolution of engineering materials................................... 8 2.3 Comparison of different unit cell stacking and............. 8 their influence on the relative strength 2.4 Deformation of 316L MHS under crush test................. 9 2.5 Temperature distribution inside the MHSS........................ 10 2.6 Mode shapes gained by numerical simulation.................... 10 2.7 a) Connection of single spheres; ...................................... 11 b) Multi-sphere network 2.8 Simplified models for sintered hollow sphere..................... 13 structures: a) Flattened contact region; b) Point contact; c) Syntactic hollow sphere structure (new development) 2.9 Simplified models of PHSS: a) Non-perforated shell; ...... b) Perforation allowing the matrix to fill the cavity; c) PHSS fully embedded in the matrix 14 xii 2.10 Single hollow spheres: a) closed surface (common.............. 14 configuration); b) with perforated surface (new development); (© by Glatt GmbH, Dresden, Germany) 2.11 Modeling of structures by gathering discrete elements......... 18 (a) One-dimensional elements. (b) Two-dimensional plane stress elements 2.12 Representative volume element of a simple cubic cell......... 25 Structure 3.1 Flowchart of the research methodology................................ 22 3.2 The derivation the one-eighth of a unit cell used in the........ 23 Finite Element Analysis (FEA) package 3.3 Schematic diagram for the PHSS embedded in the matrix:.. 24 a) PHSS and matrix with distance of adhesively bonded necking, d (i.e. 0.12 mm); b) PHSS and matrix with double distance of adhesively bonded necking, 2d; c) Flattened contact area of PHSS 3.4 PHSS composite exploded view: a) Inner matrix;................ 25 b) PHSS; c) Plate-shaped matrix; d) Outer matrix 3.5 Schematic illustration of primitive cubic sphere.................. 26 arrangement of perforated hollow spheres: a) bonded syntactic perforated hollow sphere structure (PHSS); b) sintered syntactic PHSS (on the right of each schematic representation, the arrangement of the spheres in 2D is shown to clarify the distance between adjacent spheres) 3.6 Front view of PHSS............................................................... 28 3.7 Front view of inner matrix..................................................... 29 3.8 Front view of outer matrix..................................................... 29 3.9 Isometric view of plate-shaped matrix...................................30 xiii 3.10 Finite element mesh of a perforated hollow sphere:.............. 31 a) arrangement with completely immersed sphere in the matrix; b) sintered arrangement (the matrix mesh is not shown for clarity) 3.11 Configuration of different hole diameters for bond.............. 32 gap, a (refer to Figure3.17) starting with arrangement of the PHSS with the smallest hole on the shell surface, hole diameter is increased from; a) HSS (non-perforated configuration) by b) PHSS with hole diameter equal to 25% of dl, c) PHSS with hole diameter equal to 50% of dl, d) PHSS with hole diameter equal to 75% of dl and e) PHSS with hole diameter equal to dl = 1.36 mm 3.12 Configuration of different hole diameters for bond............... 33 gap, 2a starting with arrangement of the PHSS with the smallest hole on the shell surface, hole diameter is increased from; a) HSS (non-perforated configuration) by b) PHSS with hole diameter equal to 25% of dl, c) PHSS with hole diameter equal to 50% of dl, d) PHSS with hole diameter equal to 75% of dl and e) PHSS with hole diameter equal to dl = 1.36 mm 3.13 Configuration of different hole diameters for sintered......... 34 starting with arrangement of the PHSS with the smallest hole on the shell surface, hole diameter is increased from; a) HSS (non-perforated configuration) by b) PHSS with hole diameter equal to 25% of dl, c) PHSS with hole diameter equal to 50% of dl, d) PHSS with hole diameter equal to 75% of dl and e) PHSS with hole diameter equal to dl = 1.36 mm 3.12 PHSS shell transformed from solid to three-dimensional.... 35 Hex-meshing xiv 3.15 Results of convergence study on mesh refinement; a) the... 36 graph shows that the value of the thermal conductivity became stable when the number of nodes in the meshed model exceeds 9000 nodes b) on left is the one-eighth of a simple cubic arrangement model with less number of nodes and on the right is the model with finer mesh 3.16 Results of convergence study on mesh refinement for the.... 37 sintered arrangement i.e. using automatically generated tetrahedral mesh; a) the graph shows that the value of the Young’s modulus became stable when the number of nodes in the meshed model reaches 14000 nodes b) on left is the one-eighth of a primitive cubic arrangement model with less number of nodes and on the right is the model with finer mesh 3.17 Finite element mesh and applied boundary conditions of a... 39 syntactic perforated hollow sphere cubic unit cell: a) arrangement with completely immersed sphere in the matrix; b) sintered arrangement (the darker grey elements belong to the sphere while the lighter grey elements belong to the matrix 3.18 Finite element mesh and applied boundary conditions........ 40 of a syntactic perforated hollow sphere cubic unit cell: a) arrangement with completely immersed sphere in the matrix; b) sintered arrangement (the darker grey elements belong to the sphere while the lighter grey elements belong to the matrix) 3.19 Schematic diagram to indicate the difference between......... 41 perforated and empty shell syntactic hollow sphere structures: a) perforated sphere with bond gap, b) empty sphere with bond gap, c) sintered perforated sphere, and d) sintered empty sphere xv 3.20 a) Micrograph of a completely single hollow sphere;......... 47 b) Micrograph with several hollow spheres; c) Micrograph of the wall of a hollow sphere (the micrographs are recorded and analysed with the Zeiss AxioVision 4.6.3. Sp1 Software in Aalan Univ. Germany). 3.21 Schematics stress-strain diagram showing linear elastic...... 44 deformation for loading and unloading cycles 3.22 The graphical interpretation of the Poisson’s ratio.............. 46 3.23 Original and deformed model with bond gap arrangement.. 47 3.24 The heat transfer principles interpreted into electrical......... 48 circuits 4.1 Young’s modulus of syntactic PHSS versus fraction of ...... 53 hole diameter: a) Using steel (AISI 8000) as the material of the shell (as a function of the hole diameter, a fraction of 0 denotes to the shell without holes, with a hole diameter of d = 1.36 mm) and embedded in epoxy matrix (a = 0.12 mm); b) Using aluminium as the material of the shell and embedded in epoxy matrix. 4.2 Poisson’s ratio of syntactic PHSS versus fraction of............ 54 hole diameter: a) Using steel (AISI 8000) as the material of the shell (as a function of the hole diameter, a fraction of 0 denotes to the shell without holes, with a hole diameter of d = 1.36 mm) and embedded in epoxy matrix (a = 0.12 mm); b) Using aluminium as the material of the shell and embedded in epoxy matrix 4.3 Young’s modulus of syntactic PHSS versus average........... 55 density: a) Using steel (AISI 8000) as the material of the shell embedded in epoxy matrix (a = 0.12 mm); b) Using aluminium as the material of the shell and embedded in epoxy matrix xvi 4.4 Poisson’s ratio of syntactic PHSS versus average................ 56 density: a) Using steel (AISI 8000) as the material of the shell embedded in epoxy matrix (a = 0.12 mm); b) Using aluminium as the material of the shell and embedded in epoxy matrix 4.5 Elastic properties dependence on the loading direction: .... 57 (a) partially-bonded MHSS, (b) syntactic MHSS 4.6 Numerical compression computation results of .................. 58 the closed-cell Alporas® aluminium 4.7 Young’s modulus of PHSS as a function of the average....... 58 density: (a) primitive cubic arrangement with links; (b) sintered primitive cubic arrangement 4.8 Thermal conductivity of syntactic PHSS versus fraction of.. 62 hole diameter: a) bonded configuration (a hole diameter fraction of 0 corresponds to the shell without holes, a = 0.12 mm); b) sintered configuration 4.9 Thermal conductivity of syntactic PHSS versus average..... 63 density: a) bonded configuration (a hole diameter fraction of 0 corresponds to the shell without holes, a = 0.12 mm); b) sintered configuration 4.10 Schematic thermal distribution inside the one-eighth........... 64 of the unit cell of bonded and sintered arrangement; a) Unit cell with bond gap; b) Sintered unit cell 5.1 The common metallic crystal structures................................ 68 5.2 Table of a milling machine made from hollow.................. 69 sphere composite, steel plate and carbon laminates 5.3 Finite element models and robot arms made from............. (a) aluminium alloy and (b) hollow sphere composite 69 xvii LIST OF ABBREVIATIONS CFRP - Carbon fiber reinforced plastic MHSS - Metallic hollow sphere structures FCC - Face-centered cubic BCC - Body-centered PC - Primitive cubic Hex - Hexagonal SPHB - Split Hopkinson Pressure Bar HSS Hollow sphere structures - PHSS - Perforated hollow sphere structures SSP Spherical sphere structures - MHSC - Metallic hollow sphere composites LMC - Lattice Monte Carlo RVE - Representative Volume Element UCs - Unit cells ave - Average ma - Matrix eff - Effective sp - Sphere xviii Ep - Epoxy resin FEM - Finite element method Al - Aluminium St - Steel AISI - American Iron and Steel Institute TPS - Transient Plane Source CT - Computed tomography xviii LIST OF SYMBOLS Latin minuscules a - Bond gap (the thickness of epoxy matrix between the spheres) bs - the radius of sintered contact area dl - Perforation hole diameter ε - Strain εtrans - Transverse strain εaxial - Axial strain ∆x - Distance between two surfaces σ - Total reaction stress from the applied displacement (MPa) l - Length of specimen ∆l - Applied displacement (boundary condition) or displacement obtained from the applied load ∆lx - Displacement resulted from the applied load in x-direction k - Thermal conductivity ri - Shell inner radius rs - Shell outer radius ro - Shell outer radius t - Shell thickness υ - Poisson’s ratio xix x - Coordinate in the unit space ρave - Average density ρrel - Relative Density ρso - Density of the sphere shell material xx Latin capitals ̇ - Total heat flux - Convective heat flux ̇ - Conductive heat flux A - Area for heat flux calculation E - Young’s modulus (MPa) K - Kelvin GB - Gigabyte ̇ RAM - Random access memory T1 , T2 - Constant temperature boundary conditions V - Volume Vfree - Total volume of the void(s) inside the unit cell Vma - Total volume of the matrix inside the unit cell Vrel - Relative volume of the voids Vso - Total volume of the solid material inside the unit cell Vsp - Total volume of the spherical shell(s) inside the unit cell VUC - Volume of the unit cell xxi LIST OF APPENDICES APPENDIX TITLE PAGE A Sample Calculation of the Volume Fraction of the Matrix....... 78 B Sample Calculation of the average density for composite ....... 81 Material 1 CHAPTER 1 INTRODUCTION 1.1 Introduction The idea of artificial cellular and porous materials originated from nature which creates structural optimization with respect to weight and load-carrying capacities. Bones, cork, wood, honeycombs and foams are natural materials to name a few, structured to have the wonderful properties according to their needs. Due to their unique cellular structure, for years people have been working on the development of artificial cellular materials in order to fulfill the potential materials demand in the near future. Starting in 1960s, the geometry of honeycombs was identically converted into aluminium structures as cores of lightweight sandwich panels in the aviation and space industries [1]. In 1970, the concept of porous and cellular metals first emerged [2-4]. The combination of specific mechanical and physical properties in the cellular materials makes the newfound composite varying from the ordinary dense metal. Cellular metals are being thoroughly investigated since they have a wide range of different possible arrangements and forms of cell structures. Open- and closed-type classical metal foams were illustrated in Figure 1.1 taken from literature [5-6]. 2 a) b) c) Figure 1.1: Cellular metals: a) M-Pore® (aluminium sponge); b) Alporas® (aluminium foam); c) Brass foam [5-6] . The usage of composite materials in various industries including marine, aerospace and chemical process plant shows that this alternative material is capable to replace traditional ferrous materials. Composite materials comprise of the reinforced phase bounded within a matrix or binder, e.g. Carbon Fiber Reinforced Plastic (CFRP) and Fibre glass. There are various reinforcing materials in terms of shape such as fibers, whiskers, cloth, braids, dispersed particles, and flakes [7-9]. For this research project, the characteristic of hollow spheres immersed in a polymer matrix was investigated. 1.2 Problem Identification Classical engineering materials utilized in many industrial fields reach their limitations in properties thus, new developments are required. The increasing demands can be satisfied in many fields with introducing advanced structured materials. For instance, syntactic foams are of a promising candidate in this context. The prediction and optimization of physical properties require the development of accurate and justified computational models from which constitutive equations and material properties can be derived. By means of an advanced commercial finite element analysis code, this research has comprehensively investigated the trend and behavior of hollow sphere structure composites based on perforated inclusions. 3 1.3 Objective The primary objective of this thesis is to develop adequate computational models based on different unit cell approaches. Optimized meshes should be determined based on mesh refinement analysis. The following physical parameters should be predicted for different geometrical properties and material sets; 1.4 i. Average mechanical properties (i.e. elastic properties) and ii. Average heat transfer properties (i.e. heat conductivity). Scope of Study The scope of this research is as follows: i. Generate finite element models for the hollow sphere composites; ii. Run simulations for different parameters; iii. Evaluation and interpretation of the numerical results. 2010 Jan Feb Mar Apr May Jun Jul Aug Nov Oct Nov Dec Jan Literature review and background study Modeling & meshing 1st 2nd 3rd c ration Preliminary simulation Full Analysis Results and discussion Writing and publishing Figure 1.2: Gantt chart 2011 Feb Mar Apr May Jun 5 1.6 Summary This chapter introduces the past and current development on hollow sphere structures. Initiating with successfully transformed natural honeycombs geometry with aluminium core, the investigation on the advanced materials continues rapidly with the novel PHSS. The shell of the HSS with the perforated structure offers a variety of specific mechanical and physical properties to be explored. The scopes and objective of this research were also highlighted in this chapter. Last but not least, the Gantt chart for this thesis was also included. 6 CHAPTER 2 LITERATURE REVIEW 2.1 Introduction This chapter covers the literature review on perforated hollow sphere structures which being the main focus in this thesis. Latest development of porous and cellular materials by a handful of researchers was also highlighted in this chapter. The knowledge of the literature review of PHSS, which covers the basic of porous structures, will give the reader deeper understanding of the following chapters. 2.2 Prospective for Lightweight Cellular Metals Cellular metals demonstrate a variety of outstanding specific mechanical and physical properties such as energy absorption [10] and high specific stiffness [11]. In conjunction with their specific properties, these cellular materials have considerable potential for a vast usage in the future. Taking metallic hollow sphere structures (MHSS) into practice in automotive industry, filler of hollow profiles to absorb the impact energy in the car front bumper [12] for instance, engineers and designers are also aiming for reduction in weight of the vehicle. Reduction in weight will bring the reduction in fuel consumption and improvement of the 7 driving dynamic hence increasing in passive safety which derives from the energy absorption capacity at low stress levels. Furthermore, the driving comfort can be increased by acoustic and thermal insulation in conjunction with the damping of driving vibrations [13]. Cellular metals are also popular due to their high strength-to-weight ratio and their high gas permeability combined with a high conductivity [1]. The Venn’s diagram, (cf. Figure 2.1) shows the summary of the benefits of cellular metals. High specific stiffness and strength Structural vibration, Acoustic damping Cellular Metals Good Energy Absorbers Good Thermal Conductors Figure 2.1: Properties of cellular metals (adopted from [14]) 2.2.1 High specific stiffness and strength Cellular metals are engineered to meet the demand of the high-end technology usage i.e. aerospace, Formula-1 and aero plane. The main criterion of the above mentioned industries is to have light yet strong material in order to optimize their need. The evolution of engineering material from the beginning is summarized in Figure 2.2 [15]. The classical metals experienced a decrease in importance while a rise of advanced composites started by late 1950 and kept improved by various scientists and engineers up to date. Sanders and 8 Gibson (2003) addressed that face-centered cubic(FCC) stacking hollow sphere structures is more superior than body-centered (BCC) and primitive cubic (PC) packing in terms of elastic properties [16]. The following diagram (cf. Figure 2.3) summarises the results. Figure 2.2: Evolution of engineering materials (adopted from [15]) Figure 2.3: Comparison of different unit cell stacking and their influence on the relative strength (adopted from [16]) 9 2.2.2 Good Energy Absorbers A study by using the crush tests on 316 L metallic hollow sphere structures have been conducted by Taşdemirci (2010). The following snapshot, Figure 2.4 from the test by using Split Hopkinson Pressure Bar (SPHB) shows how cellular materials are being able to deform up to three-fourth from their original size [17]. It is evident that cellular metals are good energy absorbers, which can be used to enhance the car safety for both driver and passengers. Figure 2.4: Deformation of 316L MHS under crush test (adopted from [17]) 2.2.3 Good Thermal Conductors Fiedler et al. (2008) conducted experimental and numerical analyses of thermal conductivity behavior in adhesively bonded metallic hollow sphere and concluded that the adhesive in MHSS acts as a good insulator [18]. The temperature distribution in the bonded and syntactic configuration illustrated in Figure 2.5 indicates that the heat distribution in the metallic shell is approximately constant. 10 Figure 2.5: Temperature distribution inside the MHSS (adopted from [18]) 2.2.4 Structural Vibration Low density property found in most cellular materials makes the composite suitable to fit in the component where excessive accelerations occur. Merkel and colleagues (2009) [19] addressed the vibroacoustic behavior of hollow sphere structure numerically and experimentally and found a good correlation between both approaches. The mode shapes of the finite element analysis models are shown in Figure 2.6. Baumeister and Molitor (2009) [12] had demonstrated the potential usage of metallic hollow sphere composite in industrial machine parts e.g.: industrial robot arm. In addition, cellular materials are also able to restrain unnecessary sound and act as an acoustic insulator. 11 Figure 2.6: Mode shapes gained by numerical simulation (adopted from [19]) 2.3 Perforated Hollow Sphere Structures The main idea of introducing holes on the sphere shells is to make the inner sphere and volume functional [20]. With the perforation in the shells, the matrix used for the composite will not linger around the shell anymore but, the matter can now fill the inner part of the sphere thus automatically change the physical and also the specific mechanical properties of the compound. Instead of using HSS with air cavities as the reinforcing particles in the matrix, now it is replaced by the perforated hollow sphere with a matrix as a filler. The connection between the spheres can be simplified as shown in Figure 2.7(a). The formation that occurred between the spheres and the matrix determines the properties on the macroscopic level [21]. 12 a) b) Figure 2.7: a) Connection of single spheres; b) Multi-sphere network (adopted from [21]) The schematic diagram shown in Figure 2.7(b) explains how the hollow spheres can be joined together to produce composite materials. There are three main procedures of joining the spheres without perforation, namely sintering, neck bonding and infiltrating the spheres. Sintering of metallic precursor spheres by applying heat and pressure is the first method [21]. Examples of sintered metal hollow spheres for titanium and 316L [22-23] are stated in the references. The contact area due to the sintering process is to a certain extent flattened or even curved resulting in difficulties to generate the computer models. Secondly, the spheres can be bonded by using a liquid phase which forms a neck region. However, the minimum distance between the two spheres does not need to be reduced to completely to zero. In [24], a minimum distance between 0.12 and 0.13 mm is reported for adhesively bonded spheres of diameter 1.0 mm. The third method of joining the spheres is to infiltrate the interstitial spaces in a three-dimensional network of hollow spheres with a liquid resin such as epoxy or a low melting temperature metal [21]. The so-called ‘syntactic’ foams are produced since the spheres are ‘constructed’, as opposed to the normal foaming process which results in a chaotic arrangement of cavities in a single material [25]. In this research project, the third method of bonding the metal spheres was taken into account since the main objective of the research is to investigate how the inclusions of the matrix in the sphere cavity affect the properties of the cellular metals. It must be noticed that the final arrangement of the spheres particles was not only periodically arranged as illustrated schematically in Figure 2.8(c), since in real life the array of the spheres will be randomly dispersed. Hence, the flattened region and point contact cases is also 13 possible to occur for the PHSS as shown in Figure 2.8(a, b). Figure 2.9 illustrates the transformation of HSS based on perforated inclusions as the matrix ideally embedded inside the spheres. The difference between the shell surface of HSS and PHSS is illustrated in Figure 2.10. a) b) c) Figure 2.8: Simplified models for sintered hollow sphere structures: a) Flattened contact region; b) Point contact; c) Syntactic hollow sphere structure (new development) (adopted from [21]) 14 b) a) c) Figure 2.9: Simplified models of PHSS: a) Non-perforated shell; b) Perforation allowing the matrix to fill the cavity; c) PHSS fully embedded in the matrix (adopted from [21]) a) b) Figure 2.10: Single hollow spheres: a) closed surface (common configuration); b) with perforated surface (new development); (© by Glatt GmbH, Dresden, Germany) (adopted from [20]) 15 2.4 Literature Assessment Cellular metals demonstrate a variety of outstanding specific mechanical and physical properties. In conjunction with their special properties, these cellular materials have considerable potential for the vast usage in the future. Furthermore, driving comfort can be improved by acoustic and thermal insulation in conjunction with the damping of driving vibrations [26]. In addition, the cellular metals are also desirable due to their high strength-toweight ratio [27]. Several techniques have been discovered since a century ago in order to manufacture hollow sphere structures [27]. In recent years, many researchers and engineers carried out experiments and simulations to come out with low-cost yet viable method to produce hollow sphere structures in bulk. Palmer et al. (2007) [28] concentrated on the pressure infiltration technique producing a low-density material with high energy absorption. Augustin and colleague (2009) [29] addressed the so-called “green spheres” to produce hollow spheres (HS) and hollow sphere structures (HSS) involving polystyrene core, coating and binder as the key materials. In the early times, the so-called spherical sphere structures (SSP) were used as main reinforcements for polymers. Muliana and Kim (2007) [30] predicted nonlinear viscoelastic responses of composites reinforced with solid spherical particles by using silicon carbide embedded in aluminium matrix. Zhang and Ma (2009) [31] proposed the characterisation of syntactic carbon foam containing hollow carbon microspheres in phenolic resin. While Dudina and co-workers (2009) [32] have reported that their work on adding metallic glass into Mg alloy shows in increased mechanical strength without critical loss of ductility instead of Mg alloy alone. The researchers continue to discover more opportunities within the attractive hollow sphere structures, where now the spheres are connected together rather than mixed in the matrix. Öchsner and Fiedler (2009) [21] discussed two configurations of the spheres connection. The connection of the spheres can be either bonded with a neck region or in syntactic formation (the spheres arrangement constructed inside the matrix). This work also 16 described three main procedures of joining the single hollow spheres; sintering, bond-neck and infiltrating. Starting with investigations on the mechanical properties, Sanders and Gibson (2003) [16] have reported thoroughly about the mechanics of body-centered cubic (BCC) and face-centered cubic (FCC). Baumeister and Molitor (2009) [12] used two different cold hardening epoxide resin systems in preparation of metallic hollow sphere composite (MHSC) to inspect the Young’s modulus and the tensile strength. Gasser and colleagues (2003 and 2004) [33-35] conducted both experimental and numerical analyses on elastic and plastic criterion of regular and periodic stacking of hollow spheres. Speich et al. (2010) [36] investigated the large plastic deformation behavior of metallic hollow sphere structures (MHSS) based on the finite element method calibrated with the experimental results. Later, Fiedler and Öchsner (2008) [37] described the mechanical properties of adhesively bonded metallic hollow sphere structures (MHSS). Macroscopic properties and damage mechanisms of stainless steel hollow sphere foam followed by X-ray tomography have been carried out by Lhuissier and colleagues (2009) [38]. Taşdemirci et al. (2010) [17] studied the compression behavior of a 316 L metallic hollow sphere (MHS) structure both numerically and experimentally. The experiment conducted by using a Split Hopkinson Pressure Bar (SHPB) test apparatus has shown very similar crushing characteristics with numerical results. Veyhl et al. (2011) [39] addressed qualitatively similar characteristics with quantitative differences of the macroscopic mechanical properties of the open-cell between M-Pore® sponge and closed cell Alporas® foam. Hollow sphere structures undergone a rework, where Hosseini et al. (2011) [40] introduced perforation holes in order to open the inner sphere volume and surface. Then, the numerical analyses of the initial yield surface were conducted on the so-called perforated hollow sphere structures (PHSS) with primitive cubic pattern. The discovery of properties for modern engineering materials is not only done on basic mechanical properties, but scientists also look into thermal properties where the product is expected to be used. For this, Elomari and co-workers (1998) [41] used aluminum-matrix composites containing thermally oxidized SiC particles and found out that the effect of particle size is quite obvious in the pressure-infiltrated composites: the larger the particles, the greater the thermal expansion of the composite. Farnsworth et al. (2010) [42] reported that the Lattice Monte Carlo (LMC) can be used to analyze multiphase materials and the study was on thermal diffusion in syntactic hollow-sphere structures. Fiedler et al. (2008) [18, 43] studied the thermal conductivity of random and metallic hollow sphere structures by 17 using the Lattice Monte Carlo (LMC) approach for the numerical study and the Transient Plane Source method (TPS) is applied in order to perform thermal measurements on experimental samples. They reported that a good agreement of the findings is observed between numerical and experimental analyses. Solórzano et al. (2009) [44] conducted a comparative study between experimental and numerical findings on metallic hollow sphere structures (MHSS). Hosseini and colleagues (2009) [14] predicted the thermal conductivity of perforated hollow sphere structures (PHSS). Fiedler et al. (2009) [45] extended the previous investigations by giving special focus to the influence of the sintered contact area and the relative density of PHSS. Besides the common properties investigation that has been done on hollow sphere structures, there are a few investigations which address another angle of exploring this ‘new type’ composite on other fields. Merkel et al. (2009) [19] conducted a comprehensive numerical and experimental investigation on vibration and acoustic behaviour of hollow sphere structures. Vesenjak and colleagues (2009) [46] numerically investigated the dynamic behaviour of metallic hollow sphere structures and reported that syntactic structures exhibit distinctly higher value of stiffness and stress in comparison to partial MHSS. While Winkler et al. (2010) [47] investigated the vibration behavior of adhesively bonded MHSS plates. Winkler et al. (2010) then conducted an acoustic study regarding the capability of the metallic hollow sphere structures to bear the sound using the time delay between two signals and on eigenmode analysis method. Their experimental and numerical simulations gave 0.1% difference for sound velocity. In recent companion studies, mechanical and thermal studies have been carried out on perforated hollow sphere structures without matrix. Here, special focus was given on the influence of the matrix and the relative density on the thermal properties. In the scope of this work, three-dimensional finite element analysis was used in order to investigate mechanical properties and the effective thermal conductivity of primitive cubic (PC) unit cell models of PHSS embedded in the matrix (i.e. epoxy) and the obtained values were compared to structures without holes. 18 2.5 Finite Element Method The term ‘finite element method’ was used for the very first time in 1960 [48], however the origin of the approach is credited to the work of Castigliano, Maxwell, Navier, Ritz and others in 19th century. The foundation of the modern finite element method is and being simultaneously brought by two groups; Argyris and co-workers (1990) [48] in Europe and Clough et al. (1965) [49] in USA. Figure 2.11 interpreted by Clough (1965) shows that finite element method in the beginning can solve the structural beam and truss elements and the method can be further extended by using two- and three-dimensional elements [50]. Figure 2.11: Modeling of structures by gathering discrete elements. (a) One-dimensional elements. (b) Two-dimensional plane stress elements (adopted from [15]) Nowadays, the finite element method has been approved as a practical engineering tool and with the help of fast-growing computer-aided technologies, huge numbers of finite element codes have been generated to simulate the real structure response. Computer simulation can open the new field of applications of PHSS based on matrix inclusions by predicting the properties. Simulating the structures numerically is cost-effective and of course can give us more informative knowledge about the possible abilities of PHSS and its applications, even before the real specimens become available. 19 2.5.1 Representative Modeling of Cellular Materials In the case of cellular materials, the influence of edge effects at the free surface of the specimen to characterize macroscopic values for the huge numbers of cells in the real arrangement can be avoided by using the “representative volume element” (RVE) simplification [51]. A RVE of a unit cell of particular porous structures can represent the whole specimen in order to determine the mechanical properties and their mathematical characterization. In this approach, a RVE will consist of a minimum amount of unit cells (UCs) as illustrated in Figure 2.12. By using a unit cell representation for the model, if the normal strains in the same principal material system applied, the common relationships for isotropic solids will remain applicable [52]. However, the precise values of the material parameters cannot be gained by this particular approach since the dependency on the pore arrangement (i.e. another possibility of lattice configuration such as FCC, BCC and Hex) and on the direction of the loads applied (i.e. tilted to the x- and y-axes) exists. This modelling technique has been successfully applied on several investigations for instance, the plastic behaviour of porous metals explanation (Green, 1972)[53], for the numerical analysis of the deformation delocalization in metallic foams (Meguid, 2002) [54] and on the elasto-plastic behaviour of porous metals (Öchsner, 2009) [55]. Figure 2.12: Representative volume element of a simple cubic cell structure (adopted from [55]) 20 2.6 Summary This chapter elaborates on the literature review of perforated hollow sphere structures and its prospective properties lie within the interesting behavior. An overview of PHSS manufacturing and other fields of investigation (i.e. acoustic and dynamic) out of mechanical and thermal properties was addressed. The potential of the cellular metals to replace traditional dense metal in the future in terms of their specific mechanical and outstanding physical properties. The birth of the idea for PHSS was also elaborated and the possible connection of the hollow spheres was discussed. In addition, this chapter also elaborates on the fundamentals of mechanical and thermal properties concerning their respective governing equations involved in this investigation. The deformation of the simulated specimen of PHSS is also discussed. The last part of this chapter includes the elaboration on the usage of representative volume elements (RVE) and its rationale in modeling cellular materials. CHAPTER 3 RESEARCH METHODOLOGY 3.1 Introduction This chapter explains in details about the construction of the perforated hollow sphere structures which were used in the numerical analysis software. The whole process was compressed in a flowchart illustrated in Figure 3.1. The chapter covers the modeling concept and approach, geometry, mesh, applied boundary conditions and the materials used for the composite. The research project used the commercially available finite element software MSC.Marc® (MSC Software Corporation, Santa Ana, CA, USA) and the respective subroutine was used to extract the result of the post-processing models at the macroscopic level. 3.2 Modeling of PHSS The size of the system of the equation can be reduced when the so-called unit-cell approach is chosen where the computational structure is reduced from a larger or infinite amount of randomly arranged spheres to a single unit-cell which is commonly based on 22 typical space groups as in crystallography [56]. Thus, the analysis was done on a unit cell and assumptions on the boundary conditions were made so that the nearest possible actual situation of the cell can be replaced by the computational model. START Geometry generation by using CAD Software Variation of geometrical parameters NO TRANSFER YES Finite element software: Mesh, Boundary Conditions, Material SUBROUTINE Solution for the system of equations Evaluation of results END Figure 3.1: Flowchart of the research methodology 23 By referring to the three dimensional (3D) diagram of the syntactic perforated hollow sphere structure in Figure 3.2, the simplest form of a unit cell can be drawn with symmetrical condition. As a result, an even cut of one-eighth of the entire sphere is obtained to provide simplicity in modelling and optimizing the computation time during the analysis. Figure 3.2: The derivation the one-eighth of a unit cell used in the Finite Element Analysis (FEA) package 24 Since the aim of this research was to focus on the effect of the inclusions of the matrix phase in the PHSS cavities, there are three possible configuration of the composite which was considered as shown in the schematic diagram in Figure 3.3. a) b) 2d d c) Figure 3.3: Schematic diagram for the PHSS embedded in the matrix: a) PHSS and matrix with distance of adhesively bonded neck, d (i.e. 0.12 mm); b) PHSS and matrix with double distance of adhesively bonded neck, 2d; c) Flattened contact area of PHSS (adopted from [24]) 25 Figure 3.4 shows the exploded view of a unit cell used in this thesis which consists of four parts namely; a) Inner matrix, b) Metallic sphere shell, c) Plate-shaped matrix, and d) Outer matrix. Due to the symmetrical behavior of the unit cell of the model, all the geometry, mesh, boundary conditions and the material applied were only applied on one-eighth of the entire model. The computer models generated in this project were referred to the previous research done by Hosseini et al. [20] in terms of size and parameter. The similarities in the scope of the size of the model i.e. perforation holes, hollow sphere structures and necking distance allow an easy comparison of the findings although the meshes were differently generated. Figure 3.5 illustrates the arrangement of a stack of perforated hollow spheres embedded in a matrix (i.e. epoxy). d) c) b) a) Figure 3.4: PHSS composite exploded view: a) Inner matrix; b) PHSS; c) Plate-shaped matrix; d) outer matrix 26 The model used in this work is a simplified approach which neglects detailed features (where the miniscule pores in the metal sphere wall resulted from manufacturing process) of the real specimens cf. 3.20 but allows a first quick prediction of physical properties. Figure 3.5: Schematic illustration of primitive cubic sphere arrangement of perforated hollow spheres: a) bonded syntactic perforated hollow sphere structure (PHSS); b) sintered syntactic PHSS (on the right of each schematic representation, the arrangement of the spheres in 2D is shown to clarify the distance between adjacent spheres) 27 3.3 Finite Element Approach: Geometry, Mesh, Boundary Conditions and Materials The assumed cubic symmetry of the primitive cubic arrangement allows that only one-eighth of a unit cell is modeled in the primary orientation, thus reducing the computation time. For a three-dimensional meshing, principally tetrahedral or hexahedral elements can be employed. Previous investigations [24, 57] have shown that hexahedral elements yielded superior performance. However, due to the geometric complexity, this research used an automatically generated tetrahedral mesh for sintered spheres, cf. Figure 3.8 (b) and for the unconnected sphere models, typical hexahedral meshes were generated as in Figure 3.8 (a). Preliminary investigations on the tetrahedral meshing were performed and it was found that the used meshes are in the converged region. The entire models composed of matrix and sphere consist of 29592 to 67080 elements for both configurations, with the largest hole radius of 0.68 mm. Subsequent models were generated where the size of the perforation hole was reduced to 75%, 50%, 25% and 0% (i.e. without perforation as limiting case for comparison) of the maximum radius as shown in Figure 3.11 – 3.13. Figure 3.15 shows the influence of mesh density on the calculated results for the heat conductivity. It can be seen that the thermal conductivity is converging to a stable value as the mesh density increases. Thus, as shown in Figure 3.15(a), a choice of 9000 nodes is reasonable to calculate the heat conductivity. Higher number of nodes will yield more accurate results but the calculation time will increase exponentially. In this research the highest number of nodes that is selected according to the computer with 3.00 GB RAM. Figure 3.15(b) illustrates the significant effect of mesh density of the model on the smoothness of the curves that contributes to the accuracy of the calculation. The model on the left in Figure 3.15(b) with less number of nodes has bigger corners on its curves compared to the smooth profile found on the model with denser number of nodes. Figure 3.16 shows the convergence trend of the results for the tetrahedral meshing generated in this paper for sintered configurations. The chosen number of nodes is 14000 where the error will be less than 1%. 28 3.3.1 Geometry The basic idea of constructing the computer model of the PHSS embedded in the matrix is to divide them into four major parts, illustrated in the figures will represent the matrix phase which is assumed to ideally fit in the PHSS cavity. Thus, the shape of the inner matrix, plate-shaped matrix and the outer matrix will follow the profile of the PHSS as it is. Each of the major parts then will be merged together in the commercially available finite element software MSC.Marc®. Thus it is very important to set one reference point of origin. The point of origin of (0, 0, 0) is set and is taken as a reference point in this research project to make sure the assembling process will be easy afterwards. The following diagrams (Figure 3.6-3.9) show the dismantled PHSS to clearly indicate the dimensions of the numerical model. t dl rl Figure 3.6: Front view of PHSS 29 ri Figure 3.7: Front view of inner matrix ro Figure 3.8: Front view of outer matrix 30 t Figure 3.9: Isometric view of plate-shaped matrix The physical parameters of the model generated in this research project are listed in Table 3.1. Figure 3.10 shows the finite element mesh of the sphere shell and the key dimensions to build the model. Table 3.1: Summary of the considered geometry Dimension Value [mm] ri 1.25 rl 1.35 ro 1.47 dl 1.36 bs 0.3 t 0.10 a 0.12 31 Figure 3.10: Finite element mesh of a perforated hollow sphere: a) arrangement with completely immersed sphere in the matrix; b) sintered arrangement (the matrix mesh is not shown for clarity) 32 a) b) c) d) e) Figure 3.11: Configuration of different hole diameters for bond gap, a (refer to Figure3.17) starting with arrangement of the PHSS with the smallest hole on the shell surface, hole diameter is increased from a) HSS (non-perforated configuration) by b) PHSS with hole diameter equal to 25% of dl, c) PHSS with hole diameter equal to 50% of dl, d) PHSS with hole diameter equal to 75% of dl and e) PHSS with hole diameter equal to dl = 1.36 mm 33 a) b) c) d) e) Figure 3.12: Configuration of different hole diameters for bond gap, 2a arrangement of the PHSS with the smallest hole on the shell surface, hole diameter is increased from a) HSS (non-perforated configuration) by b) PHSS with hole diameter equal to 25% of dl, c) PHSS with hole diameter equal to 50% of dl, d) PHSS with hole diameter equal to 75% of dl and e) PHSS with hole diameter equal to dl = 1.36 mm 34 a) b) c) d) e) Figure 3.13: Configuration of different hole diameters for sintered arrangement with the smallest hole on the shell surface, hole diameter is increased from a) HSS (non-perforated configuration) by b) PHSS with hole diameter equal to 25% of dl, c) PHSS with hole diameter equal to 50% of dl, d) PHSS with hole diameter equal to 75% of dl and e) PHSS with hole diameter equal to dl = 1.36 mm 35 3.3.2 Mesh The mesh of the model plays a vital role in gaining reliable results in a finite element analysis. Every part in the model should have identical yet homogeneous meshing for the best results. Otherwise, if the meshing is not consistent i.e. in terms of the location of the nodes contact points, it can mess up the entire calculation done by the computer. For the threedimensional meshing, principally tetrahedral or hexahedral elements can be employed. However, investigations [24, 57] have shown that hexahedral elements yield superior performance. Thus, for this research project, three-dimensional mapped-meshing with approach on the solid surface was used. Mapped-meshing approach may produce elements of the mesh brick-shaped throughout the model. Figure 3.14: PHSS shell transformed from solid to three-dimensional Hex-meshing To have a good mesh, several steps must be followed until the model is divided so that they have four corners on the surface. A finer mesh is desirable for perpendicular angle instead of tilted ones as it is not stable. The configurations (cf. Figure 3.14) considered oneeighth of a PHSS divided further to 12 pieces. The division process was not done randomly since it must be designed so that the symmetrical characteristic remains to reduce the time of generation of the entire mesh. It is noted that each piece in Figure 3.14 has four corners so that they can be easily mapped-meshed [58]. A convergence study has been done on the mock model of the PHSS embedded in the matrix showing that the result will be stable until 36 the number of nodes reaches 9000 and above as shown in Figure 3.15 since the difference is less than 1%. The regularly subdivided mesh can be observed in Figure 3.15 (b). Figure 3.16 illustrates the convergence study result for the mesh of tetrahedral type applied on the spheres of sintered arrangement and the Young’s modulus converged at the point of 14000 nodes. 6.0 Conductivity, k [10-4W/(mm∙ K)] a) 5.5 chosen model 5.0 4.5 4.0 3.5 3.0 2.5 Bonded, 2a St-Ep 2.0 0 10000 20000 30000 40000 50000 Number of nodes b) Figure 3.15: Results of convergence study on mesh refinement; a) the graph shows that the value of the thermal conductivity became stable when the number of nodes in the meshed model exceeds 9000 nodes b) on left is the one-eighth of a primitive cubic arrangement model with less number of nodes and on the right is the model with finer mesh 37 5500 Young's modulus (MPa) a) chosen model 5000 4500 4000 3500 Sintered St-Ep 3000 0 10000 20000 30000 40000 50000 60000 Number of nodes b) Figure 3.16: Results of convergence study on mesh refinement for the sintered arrangement i.e. using automatically generated tetrahedral mesh where the graph shows that the value of the Young’s modulus became stable when the number of nodes in the meshed model reaches 14000 nodes b) on left is the one-eighth of a primitive cubic arrangement model with less number of nodes and on the right is the model with finer mesh 38 3.3.3 Boundary Conditions All considered boundary conditions in this research project were applied only on each of the outer surface of the meshed model. The so-called periodic boundary condition was introduced since the term ‘periodic’ represents a periodic continuation of two of the model perpendicular to the xy- and xz- planes, respectively. Therefore, for these two surfaces all the nodes on their corresponding surface were tied to one ‘master’ node. This means that all nodes on this surface will have the same displacement (u = const.). Theoretically, when the node is tied together, every node on those sides will behave in similar manner or have a constant deformation with their master. The summary of the boundary conditions is demonstrated in Figure 3.17. The displacement is applied in y-direction by considering a value of 0.1 mm. The finite element models of the syntactic perforated hollow sphere structure consist of two basic configurations where a perforated hollow sphere structure (PHSS) is mixed with a matrix i.e. an epoxy matrix, in the manufacturing process. For the first configuration, spheres without directly contacting each other are arranged in a primitive cubic periodic pattern where a minimum distance (matrix gap) of length 2a between adjacent spheres occurs (due to symmetry, the length shown in Figure 3.17 (a) is equal to 1a). In addition, this first configuration was modified by the option that the minimum distance between adjacent spheres can vary and in a second set of computational models this distance was increased by a factor two. Thus, a similar arrangement as shown in Figure 3.15 (a) was used where the distance was increased to 2a. The second possible configuration consists of directly connected spheres as illustrated in Figure 3.17 (b) where flattened contact areas between the spheres were created due to a joining sintering process. Due to the symmetry of the applied load and the geometry, reflective boundary conditions (ux = uy = uz = 0) were applied at three perpendicular surfaces, i.e. along the symmetry planes of the singles sphere. The influence of neighboring cells in the primitive cubic pattern is considered by the so-called repetitive boundary conditions. Reflective boundary condition means that every node located on the described surface is constrained to not move in the direction of the surfaces’ normal vector and repetitive boundary condition means that every node located on a linked surface must experience the same displacement perpendicular to the surface [59]. The thin lines on the right side in Figure 3.17 and Figure 3.18 represent the ‘periodic’ continuation of the spheres in PHSS packing. For the heat transfer analysis, zero heat flux or 39 ( ̇ = 0) is imposed on the left and right wall of the one-eighth of a unit cell and T1 and T2 on the bottom and top surfaces is set to 300K and 320K respectively. Figure 3.17: Finite element mesh and applied boundary conditions of a syntactic perforated hollow sphere cubic unit cell: a) arrangement with completely immersed sphere in the matrix; b) sintered arrangement (the darker grey elements belong to the sphere while the lighter grey elements belong to the matrix) 40 Figure 3.18: Finite element mesh and applied boundary conditions of a syntactic perforated hollow sphere cubic unit cell: a) arrangement with completely immersed sphere in the matrix; b) sintered arrangement (the darker grey elements belong to the sphere while the lighter grey elements belong to the matrix) 41 A more realistic model for the spheres without perforation holes can be obtained by assuming that the inner volume of the spheres remains empty, i.e. without any reinforcing matrix. This is schematically shown in Figure 3.19 (b) and (d). Figure 3.19: Schematic diagram to indicate the difference between perforated and empty shell syntactic hollow sphere structures: a) perforated sphere with bond gap, b) empty sphere with bond gap, c) sintered perforated sphere, and d) sintered empty sphere 42 3.3.4 Materials The materials used in this project research are the same with the one used by Hosseini et al. [60] in order to compare the findings. The summary of the materials is tabulated in Table 3.2 and Table 3.3. Table 3.2: Specific mechanical properties of the spheres and binder [13] Material Young’s Poisson’s ratio Density modulus (MPa) (-) (kg/dm3) Steel 110000 0.30 6.95 Aluminium 58820 0.34 2.67 Epoxy 2460 0.36 1.13 Table 3.3: Heat conductivity properties of base materials. [18, 20] Material Conductivity [W/(mm∙ K)] Density (kg/dm3) Steel 0.05 6.95 Aluminium 0.232 2.67 0.000214 1.13 Epoxy 43 Figure 3.20: a) Micrograph of a completely single hollow sphere; b) Micrograph with several hollow spheres; c) Micrograph of the wall of a hollow sphere (the micrographs are recorded and analysed with the Zeiss AxioVision 4.6.3.Sp1 Software in Aalan Univ. Germany). (adopted from [28]) As shown in Figure 3.20, one can see that the wall of hollow spheres is porous. This porosity is the result of vaporing of EPS (Expanded Poly Styrol) granulates from inside of hollow sphere during manufacturing. Therefore, reduced values taken from literature [40]have been used for the base materials properties. It has been shown that the behaviour of the base material in wall thickness of PHSS is isotropic [29]. 44 3.4 Mechanical Properties 3.4.1 Young’s Modulus and Poisson’s Ratio Young’s Modulus The ratio of the stress over strain is known as Young’s modulus. This is known as Hooke’s law, introduced by Robert Hooke back in 1678 [61], and the constant of proportionality E (GPa or equivalent to psi) is the Young’s modulus . Stress Unload Slope = E 1 Load Strain Figure 3.21: Schematics stress-strain diagram showing linear elastic deformation for loading and unloading cycles [30] ( ⁄ ) = = (∆ ⁄ ) Where, σ = the total reaction force from the applied displacement divided by the cross-sectional area l = the initial length of the specimen ∆l = applied displacement boundary condition F = summation of reaction forces ε = strain [3.1] 45 Poisson’s ratio The negative ratio of the fraction of contraction (strain) and the fraction of elongation (strain) is called Poisson’s ratio. The Poisson’s effect is named after the contraction and elongation ratio phenomenon. Theoretically, Poisson’s ratio for isotropic materials should be in the range of 0 to 0.5. Furthermore, the maximum value for (or that value for which there is no net volume change) is 0.50 [62]. For many metals and metal alloys, the value range is between 0.25 and 0.35 as can be found in literature [63-65]. The three-dimensional representation of a cube constructed in Figure 3.22 demonstrates the existence of transverse and axial strains whenever a metallic block is elongated. = − d d =− d d =− d d [3.2] Where, = The resulting Poisson’s ratio d = The transverse strain [negative for axial tension (stretching), positive for axial compression i.e. y and zdirection] d = The axial strain [positive for axial tension, negative for axial compression i.e. x-direction] ∆l = applied displacement 46 Figure 3.22: The graphical interpretation of the Poisson’s ratio 3.4.2 Deformation of the Model The original and deformed model is shown in Figure 3.23. The colors show the difference of the reaction force experienced by the nodes at a different location. The lighter grey region represents the higher reaction force exerted onto the area. All the models show a similar pattern of deformation thus proving the consistency of the boundary conditions applied. It can be concluded that the surface contacted with x-, y- and z-plane remains to their respective planes (symmetry boundary condition) and just elongates to y-direction since the displacement of 0.1 mm was applied in this analysis. In Figure 3.23, it is noted that there are two faces (i.e. xy- and yz- faces) contracted constantly since periodic boundary conditions were applied on the respective surfaces. 47 Figure 3.23: Original and deformed model with bond gap arrangement 3.5 Thermal Properties Thermal Conductivity Dated back to 1807, Jean Baptiste Joseph Fourier has given a contribution in formulating the transient process of heat conduction, described by a differential equation [66]. As shown in Figure 3.24, the heat can be transferred through three mechanism (i.e. thermal conduction, ̇ convection, ̇ ; and thermal radiation, ̇ ). The total heat flux, ̇ ; thermal (the rate of heat energy transfer through a given surface) can be determined in terms of the outer temperatures (i.e. T1 and T4), providing the effective thermal conductivity of the 48 system. A in equation 3.3 denotes the area of the control surface and ∆x is the distance between these surfaces. = ̇ ∆ ∙ [3.3] Figure 3.24: The heat transfer principles interpreted into electrical circuits [13] In this thesis, we only investigated the thermal conductivity of PHSS based on matrix inclusions where the significant conduction in solids is based on lattice vibration and movement of free electrons. Thus, the equation to compute the conduction of thermal energy based on Fourier’s law is as follows; ̇ =− ∙ ∆ ∆ where the parameter k is known as thermal conductivity. [3.4] 49 3.6 The Average Density for Composite Materials From the relative density equation proposed by Gibson and Ashby (1997) [67], the average density of cellular materials can be computed through following formulae. For the cellular materials, physical properties are usually described as a function of their relative density. = Where, (3.5) : The density of the cellular material : The density of the sphere shell material In literature [11], the scaling relations or power-laws for different physical properties were discussed. In this thesis, both the masses of the solid material and cellular materials is identical; the RHS of equation 3.5 is divided by the mass to state the relative density in terms of volume. The equation is also known as the relative solid volume. = Where, (3.6) : The volume of the solid material (e.g. the spheres, the bond material) : The volume of a unit cell 50 The overall volume of a unit cell in HSS in this study including the solid and free volume can be expressed as: = + (3.7) By referring to Figure 3.3, the solid volume consists of the perforated sphere shells (index ‘sp’) and the bond material (index ‘ma’). Now the total volume of a unit cell can be calculated by: = + (3.8) It must be noted on the assumption made that the densities of the sphere shell and the matrix (bond) material are identical, (i.e. ρso = ρsp = ρma) hence, the relative density is: = = + (3.9) In the case that the sphere shell material and the bond material are different (e.g. steel and epoxy), instead of relative density which is no longer possible, an average density, ρ can be used: = ∙ + ∙ (3.10) 51 3.7 Summary This chapter highlights the development of the PHSS model generated by computer aided drawing (CAD) software; its simplification and assumptions. The possible arrangement of the hollow spheres in the matrix is also explained. A set of models was generated using the commercial available finite element software MSC.Marc®. The overview of the basic criteria of FEA approach including the considered geometry, mesh, applied boundary conditions and the base materials chosen are also treated. The chapter ends up with the derivation of average density for common cellular materials. 52 CHAPTER 4 RESULTS AND DISCUSSION 4.1 Introduction The numerical model of the PHSS made of steel and aluminium spheres embedded in the epoxy matrix has undergone the numerical analysis. A set of numerical results is presented in this chapter. The results gained are compared to the previous companion work including the discussions on the trend of the graphs. The details of the findings of the simulation are discussed further in the sub-chapter. 4.2 Young’s Modulus and Poisson’s Ratio Figures 4.1 and 4.2 summarize the macroscopic Young’s modulus and Poisson’s ratio of both cases i.e. bonded and sintered arrangements of perforated hollow sphere structures for steel-epoxy and Aluminium-epoxy composites. It can be seen that Young’s modulus (cf. Figures 4.1a and 4.1b), increases as the hole diameter 53 decreases. This can be explained by the fact that the stiffer metal is more replaced by the softer epoxy matrix. 6000 a) Sintered Young's modulus (MPa) 5500 5000 Bond gap, a 4500 4000 3500 Bond gap, 2a 3000 St-Ep 2500 2000 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Fraction of hole diameter Young's modulus (MPa) b) 5000 4500 4000 3500 3000 2500 2000 1500 1000 500 0 Sintered Bond gap, a Bond gap, 2a Al-Ep 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Fraction of hole diameter Figure 4.1: Young’s modulus of syntactic PHSS versus fraction of hole diameter: a) Using steel (AISI 8000) as the material of the shell (as a function of the hole diameter, a fraction of 0 denotes to the shell without holes, with a hole diameter of d = 1.36 mm) and embedded in epoxy matrix (a = 0.12 mm); b) Using aluminium as the material of the shell and embedded in epoxy matrix. 54 0.350 a) Bond gap, 2a Sintered Poisson's ratio 0.345 0.340 Bond gap, a 0.335 0.330 0.325 St-Ep 0.320 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Fraction of hole diameter 0.355 b) Poisson's ratio 0.350 Sintered 0.345 Bond gap, 2a 0.340 Bond gap, a 0.335 0.330 0.325 Al-Ep 0.320 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Fraction of hole diameter Figure 4.2: Poisson’s ratio of syntactic PHSS versus fraction of hole diameter: a) Using steel (AISI 8000) as the material of the shell (as a function of the hole diameter, a fraction of 0 denotes to the shell without holes, with a hole diameter of d = 1.36 mm) and embedded in epoxy matrix (a = 0.12 mm); b) Using aluminium as the material of the shell and embedded in epoxy matrix 55 6000 a) Sintered Bond gap, a Bond gap, 2a Young's modulus (MPa) 5500 5000 4500 4000 3500 3000 St-Ep 2500 2000 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 Average density (kg/dm3) 4500 Young's modulus (MPa) b) Sintered Bond gap, a Bond gap, 2a 4000 3500 3000 2500 Al-Ep 2000 1.16 1.18 1.20 1.22 1.24 1.26 1.28 1.30 1.32 Average density (kg/dm3) Figure 4.3: Young’s modulus of syntactic PHSS versus average density: a) Using steel (AISI 8000) as the material of the shell embedded in epoxy matrix (a = 0.12 mm); b) Using aluminium as the material of the shell and embedded in epoxy matrix 56 0.350 a) Poisson's ratio 0.345 0.340 0.335 0.330 Sintered Bond gap, a Bond gap, 2a 0.325 St-Ep 0.320 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 Average density (kg/dm3) b) 0.355 Poisson's ratio 0.350 0.345 0.340 Sintered Bond gap, a Bond gap, 2a 0.335 Al-Ep 0.330 1.15 1.20 1.25 1.30 1.35 Average density (kg/dm3) Figure 4.4: Poisson’s ratio of syntactic PHSS versus average density: a) Using steel (AISI 8000) as the material of the shell embedded in epoxy matrix (a = 0.12 mm); b) Using aluminium as the material of the shell and embedded in epoxy matrix 57 Comparing the both bond gap and sintered configurations, it is evident that the directly connected sphere arrangement is clearly stiffer than the spheres with a bond gap. Furthermore, the entire stiffness of the structure decreases as the bond gap increases (bond gap a to 2a). The pattern of the graphs plotted (cf. Figure 4.1) correlates well with the findings in [37, 39, 59] in terms of the trend as shown in the following Figures 4.5, 4.6 and 4.7. Figure 4.5: Elastic properties dependence on the loading direction: (a) partially-bonded MHSS, (b) syntactic MHSS (adopted from [37]) 58 Figure 4.6: Numerical compression computation results of the closed-cell Alporas® aluminium foam (adopted from [39]) Figure 4.7: Young’s modulus of PHSS as a function of the average density: (a) primitive cubic arrangement with links; (b) sintered primitive cubic arrangement (adopted from [60]) 59 The trend of Poisson’s ratio as shown in Figure 4.2 is opposed to the general behavior of Young’s modulus. As such, an increase of the hole diameters increases Poisson’s ratio. This comes from the fact that with increasing hole diameter more matrix is accumulated in the unit cell. The trend of the results obtained follows the study conducted by Hosseini et al. [37, 59] on ‘neck bonded’ PHSS with primitive cubic arrangement. In both diagrams (cf. Figures 4.1 and 4.2), the behavior of sintered and bonded structures exhibits a linear tendency. The filled markers on the vertical axis are shown to indicate the results of the sphere shells with no matrix inside in order to compare with these new type configurations. The schematic diagram to show the difference between the filled and empty shell is illustrated in Figure 4.1 and Figure 4.2. The values gained from the simulation of the empty shell are relatively lower in the case of elastic modulus compared to the shell filled with the matrix. In contrary, in the case of Poisson’s ratio, the result of the empty shells is slightly higher than that of the shells accumulated with the matrix. Thus, from this observation, it can be concluded that the matrix elements clearly stiffen the perforated hollow sphere structures. Figures 4.3 and 4.4 show the influence of the average density (hole diameter) and different joining methods on Young’s modulus and Poisson’s ratio for steel-epoxy and Aluminium-epoxy combinations. It can be seen from Figures 4.3(a) and 4.3(b), the Young’s modulus increases as the average density increases (hole diameter decreases) for all considered joining techniques. Moreover, the behavior of these structures seems to be nearly continuous in terms of strength from sintered to bond gap, a and bond gap 2a arrangement. While for the Poisson’s ratio, the values decrease with increasing average density (decreasing hole diameter) (cf. Figure 4.4). The structures with bond gap arrangements behavior tend to follow a linear shape whereas the sintered structures follow some kind of linear trend with a slight different of the values. In order to avoid a discontinuity in the distribution, the spheres without a perforation hole are still assumed to be filled with the epoxy matrix. This assumption can be questioned from the point of manufacturing but results in a much smoother curve for the limiting case. For the mechanical properties, the rule of mixtures [68] of PHSS composite is applied to obtain the value of Young’s modulus, E in the beginning of this work. This calculation also useful to check whether numerical approach is needed or not since, 60 the rule of mixtures itself can be used to calculate the Young’s modulus of large particle composites. From this rule, we can calculate the value of E of the composite (Ec) in terms of elastic moduli of the matrix phase (Em) and the particulate or metallic sphere phase (Ep), by using the following formulation; = + (1 − ) [4.1] where vm is the volume fraction of the matrix phase. However, this equation is not applicable for this complex geometry composite since the Young’s modulus value obtained seems to be too far from the numerical values. 4.2 Thermal Conductivity In the following, the results of the finite element analyses are shown. Figure 4.8 summarises the macroscopic conductivity of syntactic perforated hollow sphere structures for bonded and sintered arrangements. It can be seen in Figure 4.8(a) (configurations with different bond gaps) that there is dominant influence of the matrix and the macroscopic thermal conductivity of the composite is in the same order of magnitude as the thermally isolating matrix. Different sphere materials, i.e. aluminium or steel, or the size of the bond gap, i.e. a or 2a, have only a minor influence on the macroscopic conductivity and the entire behavior is determined by the matrix material. The situation is completely opposed in the case of sintered configurations, cf. Figure 4.8(b), where relatively higher conductivity values are obtained since the heat can easily flow through the interconnected metallic spheres. For all the configurations, i.e. bonded or sintered, the same trend is obtained where the thermal conductivity decreases with increasing perforation diameter. This decrease in conductivity comes from the fact that with increasing hole diameter, more matrix is accumulated in the unit cell, thus enforcing the insulating effect. The filled markers in Figure 4.8 at a hole fraction of zero (i.e. spheres without 61 perforation) indicate the results for sphere shells with no matrix inside or in other words for ‘empty’ spheres. These models were additionally created to investigate if a filling matrix inside the hollow spheres significantly influences the heat conductivity for unperforated spheres. As can be seen in Figure 4.8, the empty (represented by solid markers) and filled unperforated spheres reveal practically the same conductivity which is a direct result of the assumed matrix, i.e. a thermal isolator. Figure 4.9 shows the same results but in a more common graphical representation in the context of composite materials. Here, the thermal conductivity is represented as a function of the average density (densities of the single components are weighted with their volume fraction to calculate the average density value). It can be seen in both diagrams of Fig 4.9 that the thermal conductivity increases now as the average density increases (i.e. hole diameter decreases). The numerical result obtained in this thesis shows a good agreement with the existing results in [20, 44]. Comparing the graph of thermal conductivity versus the hole diameter fraction and the thermal conductivity versus the average density, similar conclusions can be taken in regards to the influence of the joining technique and sphere shell material. Let us highlight at the end of this section that the two investigated configurations, i.e. where either the spheres are not in direct contact or directly connected due to a sintering process, do not give - from a practical point of view - a significant difference in the overall weight but can change the macroscopic thermal conductivity by several orders of magnitude. This offers a great opportunity to design and tailor an advanced structured material according to different requirements, i.e. thermally isolating or conducting. 62 Figure 4.8: Thermal conductivity of syntactic PHSS versus fraction of hole diameter: a) bonded configuration (a hole diameter fraction of 0 corresponds to the shell without holes, a = 0.12 mm); b) sintered configuration. 63 a) b) Figure 4.9: Thermal conductivity of syntactic PHSS versus average density: a) bonded configuration (a hole diameter fraction of 0 corresponds to the shell without holes, a = 0.12 mm); b) sintered configuration. 64 The schematic temperature distribution as shown in Figure 4.10 illustrates how the steady state temperature gradient varies inside bonded and sintered PHSS. Due to the high thermal conductivity in the metallic base material (i.e. sphere shell), the temperature surrounding the metallic shell is nearly constant for the bond gap arrangement. Meanwhile for the sintered arrangement, the temperature is also distributed within the sphere shell since a small portion of the inter-connected shell is Figure 4.10: Schematic thermal distribution inside the one-eighth of the unit cell of bonded and sintered arrangement; a) Unit cell with bond gap; b) Sintered unit cell. 65 exposed to the heat. The diagram visualizes that for bond gap arrangement, the isotherm (boundaries between the two contrast grey areas) occur with short space indicating that the adhesive (i.e. epoxy) acts as a good insulator. A similar trend was found in [18] by Fiedler et al. where the metallic hollow sphere structures used by them consists of steel and epoxy as the base materials. 4.4 Summary This chapter elaborates on the mechanical properties and the deformation behavior of the models after the considered geometry, materials and boundary conditions applied. Then, the numerical results of Young’s modulus and Poisson’s ratio of the composites are comprehensively discussed. Both steel and aluminium spheres combined with epoxy as the matrix phase give identical pattern of mechanical and thermal properties. 66 CHAPTER 5 CONCLUSION AND FUTURE STUDY 5.1 Conclusions Mechanical Properties Hollow sphere structures based on perforated inclusions have been examined numerically to determine the Young’s modulus and the Poisson’s ratio. The findings in this work were compared to the previous companion study as the dimensions and the base materials used are the same as in literature [59-60, 69]. It can be seen that from the comparison, the value of the Young’s modulus obtained in this study for all cases (i.e. bond gap, a; bond gap, 2a; and sintered arrangement) is slightly higher than what is reported by Öchsner et al. (2010). In addition, the value of the Young’s modulus is proportional to the size of the perforation hole diameter and the average density, whereas the bigger the hole the weaker the composites will be. As a conclusion, the matrix component accumulated inside the sphere via perforation holes helps to stiffen the structures. In contrast, the value of the Poisson’s ratio is lower compared to the result of PHSS without matrix inclusions (i.e. empty PHSS). Similar to the Young’s modulus, the result of the Poisson’s ratio follows the linear pattern when it is plotted against the size of the perforation hole and average density. In addition, the results gained by using different 67 combination of the base material (i.e. Steel-epoxy and aluminium-epoxy) give a similar tendency for parameters namely Young’s modulus and Poisson’s ratio. Thermal Properties The thermal conductivity of the PHSS with matrix inclusions was studied and compared to the work done by Hosseini et al. [20, 69]. It can be seen that a good agreement were shown between the both results. As a consequence, increasing in the size of the perforation hole (average density increases), the thermal conductivity of the composite decreases. It can be concluded that, the matrix component does not only bind the spheres, but also strengthen the insulating effect. 5.1 Future Study Metallic Crystal Structures As discussed with Professor Dr. –Ing. Andreas Öchsner, there is still a lot more research that can be done on the PHSS in various fields. Investigations of HSS based on perforated inclusions can be further extended to different periodic arrangements (cf. metallic crystal structures). The study of PHSS can cover different model structures taken from common crystal lattices including FCC, BCC, Hex (cf. Figure 5.1) and the real structure obtained from microscopic Computed Tomography (CT) scans. 68 PC (done) FCC BCC HEX Figure 5.1: The common metallic crystal structures PHSS Applications The lightweight and low strength to weight ratio properties found in PHSS will open a vast field of its applications. As discussed before in Chapter 2, cellular and porous materials can be used as crash absorbers, thermal isolators, lightweight constructions and also vibration absorbers in machine part with excessive accelerations. The studies conducted by Baumeister et al. (2004) have shown that metal foams can be utilized as machine building parts (i.e. robot arm and milling machine table). Many parts have been produced with good accuracy and tested with good results [70-71]. Example of the parts built by Baumeister at al. illustrated in Figures 5.2 and 5.3. 69 Figure 5.2: Table of a milling machine made from hollow sphere composite, steel plate and carbon laminates (adopted from [70]) Figure 5.3: Finite element models and robot arms made from (a) aluminium alloy and (b) hollow sphere composite (adopted from [71]) 70 Experimental Investigations The novel perforated hollow sphere structures are currently under development. So far, there is no available experimental data on PHSS with matrix inclusions to compare with the simulated findings. However, Fiedler et al. (2008) have carried out three different methodologies (i.e. experimental, numerical and analytical) on the thermal conductivity of metallic hollow sphere structures and showed a good agreement between the results gained [18]. In addition, the cost of experimental studies can be reduced by simulating the new type by numerical analysis. The numerical study could also open vast possible applications and properties prediction of PHSS via simulations where extensive information can be obtained. Nevertheless, experimental test is indispensably necessary and is still under development on PHSS. There is an excellent prospective to explore PHSS based on matrix inclusions in this particular field concerning the mechanical, thermal and material and their applications. 71 References [1] A. Öchsner and C. Augustin, "Introduction," in Multifunctional Metallic Hollow Sphere Structures: Manufacturing, Properties and Application, ed: Springer-Verlag Berlin Heidelberg, 2009, pp. 1–3. [2] N. N., "Metal foam," Composites Science and Technology, vol. 2, p. 11., 1971. [3] H. Bray, "Design opportunities with metal foam," Engineering Materials and Design, vol. 16, pp. 16-19, 1972. [4] R. R. Barton, et al., "Use of reticulated metal foam as flash-back arrestor elements" Chemical Engineer - London, vol. 291, pp. 708-708, 1974. [5] J. Banhart, "Manufacture, characterisation and application of cellular metals and metal foams," Progress in Materials Science, vol. 46, pp. 559–632, 2001. [6] F. Topin, et al., "Contents: Adv. Eng. Mater. 9/2006," Advanced Engineering Materials, vol. 8, pp. 767-772, 2006. [7] A. K. Kaw, Mechanics Composite Materials Second Edition. Boca Raton: CRC Press, Taylor & Francis Group, 2006. [8] H. Altenbach, et al., Mechanics of Composites Structure Elements. Berlin: Springer, 2004. [9] E. G. Wolff, Introduction to the Dimensional Stability of Composite Materials. Lancaster: DEStech Publications, Inc., 2004. [10] A. G. Evans, et al., ""Multifunctionality of cellular metal systems", ," Progress in Mtls. Sci., vol. 43, pp. 171-221, 1998. [11] M. F. Ashby, et al., Metal foams: a design guide vol. 23. Oxford, UK: Butterworth-Heinemann, 2000. 72 [12] E. Baumeister and M. Molitor, "Hollow Spheres in Composite Materials and Metallic Hollow Sphere Composites (MHSC)," in Multifunctional Metallic Hollow Sphere Structures: Manufacturing, Properties and Application, ed: Springer-Verlag Berlin Heidelberg, 2009, pp. 109–136. [13] T. Fiedler, "Numerical and Experimental Investigation of Hollow Sphere Structures in Sandwich Panels," Ph.D Thesis, University of Aveiro, Portugal, 2007. [14] S. M. H. Hosseini, et al., "Finite element simulation of the thermal conductivity of perforated hollow sphere structures (PHSS): Parametric study," Materials Letters, vol. 63, pp. 1135–1137, 2009. [15] W. O. Soboyejo, "Introduction to Advanced Materials," in Advanced Structural Materials: Properties, Design Optimization, and Applications, ed New York: Taylor & Francis Group, LLC, 2006. [16] W. S. Sanders and L. J. Gibson, "Mechanics of BCC and FCC hollow-sphere foams," Materials Science and Engineering A, vol. 352 pp. 150-161, 2003. [17] A. Taşdemirci, et al., "Split Hopkinson pressure bar multiple reloading and modeling of a 316 L stainless steel metallic hollow sphere structure," International Journal of Impact Engineering, vol. 37, pp. 250–259, 2010. [18] T. Fiedler, et al., "Numerical and experimental analysis of the thermal conductivity of metallic hollow sphere structures," Materials Letters, vol. 62, pp. 1204–1207, 2008. [19] M. Merkel, et al., "On the Vibroacoustic Behaviour of HSS," in Multifunctional Metallic Hollow Sphere Structures: Manufacturing, Properties and Application, ed: Springer-Verlag Berlin Heidelberg, 2009, pp. 71-88. [20] S. M. H. Hosseini, et al., "Finite element simulation of the thermal conductivity of perforated hollow sphere structures (PHSS): Parametric study," Materials Letters, vol. 63, pp. 1135–1137, 2009. [21] A. Öchsner and T. Fiedler, "Geometrical Properties of Hollow Sphere Structures," in Multifunctional Metallic Hollow Sphere Structures: Manufacturing, Properties and Applicationn, ed: Springer-Verlag Berlin Heidelberg, 2009, pp. 31–46. 73 [22] O. Anderson, et al., Novel metallic hollow sphere structures: processing and properties. In: Banhart, J., Ashby, M.F., Fleck, N.A. (eds.) Metal foams and porous metal structures. Bremen: MIT Publishing, 1999. [23] O. Anderson, et al., "Novel Metallic Hollow Sphere Structures," Advanced Engineering Materials, vol. 2, pp. 192-195, 2000. [24] T. Fiedler, et al., "Modelling the mechanical behaviour of adhesively bonded and sintered hollow-sphere structures," Mechanics of Composite Materials, vol. 42, pp. 559-570, 2006. [25] A. R. Luxmoore and D. R. J. Owen, Syntactic foams. New York: Macmillan Publishing Co., 1982. [26] T. Fiedler, "Numerical and Experimental Investigation of Hollow Sphere Structures in Sandwich Panels," Ph.D, University of Aveiro, Portugal, 2007. [27] C. Augustin and W. Hungerbach, "History and Production of Hollow Spheres," in Multifunctional Metallic Hollow Sphere Structures: Manufacturing, Properties and Application, ed: Springer-Verlag Berlin Heidelberg, 2009, pp. 5–30. [28] R. A. Palmer , et al., "Pressure infiltrated syntactic foams—Process development and mechanical properties," Materials Science and Engineering A, vol. 464, pp. 85–92, 2007. [29] C. Augustin and Wolfgang Hungerbach, "Production of hollow spheres (HS) and hollow sphere structures (HSS)," Materials Letters, vol. 63, pp. 1109– 1112, 2009. [30] A. H. Muliana and J. S. Kim, "A concurrent micromechanical model for predicting nonlinear viscoelastic responses of composites reinforced with solid spherical particles," International Journal of Solids and Structures, vol. 44, pp. 6891–6913, 2007. [31] L. Zhang and J. Ma, "Processing and characterization of syntactic carbon foams containing hollow carbon microspheres," CARBON, vol. 47, pp. 14511456., 2009. [32] D. V. Dudina, et al., "A magnesium alloy matrix composite reinforced with metallic glass," Composites Science and Technology vol. 69, pp. 2734-2736, 2009. [33] S. Gasser, et al., "Uniaxial tensile elastic properties of a regular stacking of brazed hollow spheres," Scripta Materialia, vol. 48, pp. 1617–1623, 2003. 74 [34] S. Gasser, et al., "Finite elements computation for the elastic properties of a regular stacking of hollow spheres," Materials Science and Engineering A, vol. 379, pp. 240–244, 2004. [35] S. Gasser, et al., "Microplastic yield condition for a periodic stacking of hollow spheres," Scripta Materialia, vol. 50, pp. 401–405, 2004. [36] M. Speich, et al., "Large deformation of core materials for sandwich panels. Große Verformung von Kernmaterialien für Sandwichelemente," Materialwissenschaft und Werkstofftechnik, vol. 41, pp. 270-275, 2010. [37] T. Fiedler and A. Öchsner, "On the anisotropy of adhesively bonded metallic hollow sphere structures," Scripta Materialia, vol. 58, pp. 695–698, 2008. [38] P. Lhuissier, et al., "Quasistatic mechanical behaviour of stainless steel hollow sphere foam: Macroscopic properties and damage mechanisms followed by X-ray tomography," Materials Letters, vol. 63, pp. 1113–1116, 2009. [39] C. Veyhl, et al., "Finite element analysis of the mechanical properties of cellular aluminium based on micro-computed tomography," Materials Science and Engineering A, vol. 528, pp. 4550–4555, 2011 2011. [40] S. M. H. Hosseini, et al., "Numerical investigation of the initial yield surface of perforated hollow sphere structures (PHSS) in a primitive cubic pattern," Finite Elements in Analysis and Design, vol. 47, pp. 804–811, 2011. [41] S. Elomari, et al., "Thermal Expansion Behavior of Particulate Metal-Matrix Composites," Composites Science and Technology, vol. 58, pp. 369-376, 1998. [42] J. Farnsworth, et al., "A lattice Monte Carlo analysis on thermal diffusion in syntactic hollow-sphere thermischer Diffusion structures. in ‘Lattice syntaktischen Monte Carlo’ Analyse Hohlkugelstrukturen," Materialwissenschaft und Werkstofftechnik, vol. 41, pp. 283-286, 2010. [43] T. Fiedler, et al., "Numerical analyses of the thermal conductivity of random hollow sphere structures," Materials Letters, vol. 63, pp. 1125–1127, 2008. [44] E. Solórzano, et al., "Thermal conductivity of metallic hollow sphere structures: An experimental, analytical and comparative study," Materials Letters, vol. 63, pp. 1128–1130, 2009. 75 [45] T. Fiedler, et al., "A refined finite element analysis on the thermal conductivity of perforated hollow sphere structures," Computational Materials Science, vol. 47, pp. 314–319, 2009. [46] M. Vesenjak, et al., "Dynamic Behaviour of Metallic Hollow Sphere Structures," in Multifunctional Metallic Hollow Sphere Structures: Manufacturing, Properties and Application, ed: Springer-Verlag Berlin Heidelberg, 2009, pp. 137–158. [47] R. Winkler, et al., "Finite Element Vibration Analysis of MHSS Based on 3D Tomography Image Processing," in Materials with Complex Behaviour, Advanced Structured Materials, ed Heidelberg Springer-Verlag Berlin 2010, pp. 243-256. [48] N. C. Knowles, "Finite element analysis," Computer-Aided Design, vol. 16, pp. 134-140, 1984. [49] R. W. Clough, "Original formulation of the finite element method," Finite Elements in Analysis and Design, vol. 7, pp. 89-101, 1990. [50] R. Clough, "The finite element method in structural mechanics," in Stress Analysis, O. C. Zienkiewicz and G.S. Holister, Eds., ed: John Wiley and Sons, Ltd., 1965, pp. 85-87. [51] J. Lemaitre, "A Course on Damage Mechanics," ed Berlin: Springer, 1996. [52] T. Fiedler, et al., "Structural Modeling of the Mechanical Behavior of Periodic Cellular Solids: Open-Cell Structures," Mechanics of Composite Materials, vol. 41, pp. 277-290, 2005. [53] R. J. Green, "A plasticity theory for porous solids," International Journal of Mechanical Sciences, vol. 14, pp. 215-224, 1972. [54] S. A. Meguid, et al., "FE modelling of deformation localization in metallic foams," Finite Elements in Analysis and Design, vol. 38, pp. 631-643, 2002. [55] A. Öchsner and G. Mishuris, "Modelling of the multiaxial elasto-plastic behaviour of porous metals with internal gas pressure," Finite Elements in Analysis and Design, vol. 45, pp. 104-112, 2009. [56] M. D. Graef and M. E. McHenry, Structure of Materials: An Introduction to Crystallography, Diffraction, and Symmetry. New York: Cambridge University Press, 2007. 76 [57] S. E. Benzley, et al., "A Comparison of All-Hexagonal and All-Tetrahedral Finite Element Meshes for Elastic and Elastic Plastic Analysis " presented at the In Fourth International Meshing Roundtable, Al-Buquerque, New Mexico, 1995. [58] FEMAP Version 7.0 Examples, Copyright © 1986-1999 by Enterprise Software Products Inc. [59] A. Öchsner, et al., "Numerical Simulation of the Mechanical Properties of Sintered and Bonded Perforated Hollow Sphere Structures (PHSS)," J. Mater. Sci. Technol., vol. 26, pp. 730-736, 2010. [60] S. M. H. Hosseini, "Numerical Characterization of Perforated Metallic Hollow Sphere Structures," Master, Universiti Teknologi Malaysia, Malaysia, 2009. [61] J. Rychlewski, "On Hooke's law," Journal of Applied Mathematics and Mechanics, vol. 48, pp. 303-314, 1984. [62] A. Kumar, et al., "Correlation between ultrasonic shear wave velocity and Poisson's ratio for isotropic solid materials," Acta Materialia, vol. 51, pp. 2417-2426, 2003. [63] D. C. William, Jr., "Mechanical Properties of Metals," in Materials Science and Engineering An Introduction Sixth Edition, ed: John Wiley & Sons. Inc., 2003, p. 121. [64] H. M. Lee and W.-J. Lee, "Poisson's ratio of pure metals and their nonmetallic compounds," Scripta Metallurgica et Materialia, vol. 25, pp. 965968, 1991. [65] J. P. Romain, et al., "Relation between the Grüneisen ratio and the pressure dependence of Poisson's ratio for metals," Journal of Physics and Chemistry of Solids, vol. 37, pp. 1159-1165, 1976. [66] T. Narasimhan, "Fourier’s heat conduction equation: History, influence, and connections," Journal of Earth System Science, vol. 108, pp. 117-148, 1999. [67] L. J. Gibson and M. F. Ashby, Cellular solids: structures & properties. Cambridge: Cambridge University Press, 1997. [68] E. Jacquet, et al., "Calculation of the transverse modulus of a unidirectional composite material and of the modulus of an aggregate. Application of the rule of mixtures," Composites Science and Technology, vol. 60, pp. 345-350, 2000. 77 [69] S. M. H. Hosseini, et al., "Numerical prediction of the effective thermal conductivity of perforated hollow sphere structures," Defect and Diffusion Forum vol. 283-286, pp. 6-12, 2009. [70] E. Baumeister, "Hohlkugelkomposit - Charakterisierung thermischer und mechanischer Eigenschaften eines neuen Leichtbauwerkstoffes. Hollow sphere composite - characterisation of thermal and mechanical properties of a new lightweight material " PhD thesis University of Magdeburg, Otto von Guericke, 2004. [71] E. Baumeister, et al., "Lightweight, hollow-sphere-composite (HSC) materials for mechanical engineering applications," Journal of Materials Processing Technology vol. 155–156 pp. 1839–1846, 2004. Appendix A - Sample Calculation of the Volume Fraction of the Matrix 78 The volume fraction of the matrix The matrix volume fraction of the matrix in composite material can be calculated by using the following equation [3] . Matrix volume fraction, Φm = (Φm) old < (Φm) new (Φm) 1.47mm < (Φm) new (3.1) Figure A: The schematic diagram to illustrate one-eighth of the cubic unit cell model of syntactic perforated hollow sphere structures arrangement. Sample calculation: For s = 1.47 mm To calculate the volume of one-eighth of the sphere (cf. Figure 3.19), vsp = = (3.2) ( . ) - ( . ) Appendix A - Sample Calculation of the Volume Fraction of the Matrix 79 = 1.288 mm3 – 1.023 mm3 = 0.266 mm3 To calculate the volume of one-eighth of the matrix inside the unit cell, = − = (1.47) − vm = va + vb (3.2) vm = [vcube - (vsp)] + vb (3.3) + ( . ) + ( . ) = 2.911mm3 Thus, Matrix volume fraction, (Φm)1.47 mm = . . . = 0.916 Now, we want to choose a reasonable value for new s (i.e. the new length of the unit cell). In this study, we will use the new value of s = 1.59 mm. This new value of the length of the cubic unit cell expected to be more than 10% in terms of the matrix volume fraction so that a significant different of the results will be achieved. For s = 1.59 mm the same equation is applied, vsp = = ( . ) - ( . ) Appendix A - Sample Calculation of the Volume Fraction of the Matrix 80 = 1.288 mm3 – 1.023 mm3 = 0.266 mm3 vm = va + vb = [vcube - (vsp)] + vb = − + ( . = (1.59) − ) + ( . ) = 3.755 mm3 Thus, The new matrix volume fraction, (Φm) 1.59 mm (Φm) old = 1.47 mm 0.916 = . . . = 1.182 < (Φm) new = 1.59 mm < 1.182 (which means is 18.2% more matrix than the old model with s = 1.47 mm) Hence, the length of the new cubic unit cell is chosen as equal to 1.59 mm to build the second model of the syntactic perforated hollow sphere structure configuration. Appendix B - Sample Calculation of the average density for composite material 81 The average density for composite material The average density for composite material can be calculated by using the following equation ρ= ∙ ∙ + (3.6) where, vma: Matrix Volume vsp: Sphere Volume vuc: Entire Box Sample calculation: For bond gap configuration, = 1.59 mm × 1.59 mm × 1.59 mm vuc = 4.019679 mm3 vsp = = 0.2656 mm3 = vuc – vsp vma = 4.019679 - 0.2656 = 3.754079 mm3 Thus, ∙ ρ= ρ= . . + ∙ ∙ 1.13 + . . ∙ 6.95 ρ = 1.055 + 0.459 = 1.514 kg/dm3