In this thesis, we discuss the development, implementation and testing of a piecewise linear (PWL) continuous Galerkin finite element method applied to the threedimensional diffusion equation. This discretization is particularly interesting because it discretizes the diffusion equation on an arbitrary polyhedral mesh. We implemented our method in the KULL software package being developed at Lawrence Livermore National Laboratory. This code previously utilized Palmer's method as its diffusion solver, which is a finite volume method that can produce an asymmetric coefficient matrix. We show that the PWL method produces a symmetric positive definite coefficient matrix that can be solved more efficiently, while retaining the accuracy and robustness of Palmer's method. Furthermore, we show that in most cases Palmer's method is actually a non-Galerkin PWL finite element method. Because the PWL method is a Galerkin finite element method, it has a firm theoretical background to draw from. We have shown that the PWL method is a well-posed discrete problem with a second-order convergence rate. We have also performed a simple mode analysis on the PWL method and Palmer's method to compare the accuracy of each method for a certain class of problems. Finally, we have run a series of numerical tests to uncover more properties of both the PWL method and Palmer's method. These numerical results indicate that the PWL method, partially due to its symmetric matrix, is able to solve large-scale diffusion problems very efficiently.

In polymer matrix composites (PMCs) manufacturing processes can induce de- fects, e.g., voids, fiber misalignment, irregular fiber distribution in the cross-section and broken fibers. The effects of such defects can be beneficial or deleterious de- pending on whether they cause failure suppression or enhancement by localized de- formation processes e.g., crazing, shear yielding and fiber-matrix debonding. In this study, a computational approach is formulated and implemented to develop solu- tions for general boundary-value problems for PMC microstructures that accounts for micromechanics-based constitutive relations including fine scale mechanisms of material failure. The defects considered are voids, and the microstructure is explic- itly represented by a distribution of fibers and voids embedded in a polymer matrix. Fiber is modeled as a linearly elastic material while the polymer matrix is mod- eled as an elastic-viscoplastic material. Two distinct models for the matrix behavior are implemented: (i) Drucker–Prager type Bodner model that accounts for rate and pressure-sensitivity, and (ii) improved macromolecular constitutive model that also accounts for temperature dependence, small-strain softening and large-strain harden- ing. Damage is simulated by the Gearing-Anand craze model as a reference model and by a new micromechanical craze model, developed to account for craze initiation, growth and breakdown. Critical dilatational energy density criterion is utilized to predict fiber-matrix debonding through cavitation induced matrix cracking. An extensive parametric study is conducted in which the roles of void shape, size and distribution relative to fiber in determining damage initiation and evolution are investigated under imposed temperature and strain rate conditions. Results show there are significant effects of voids on microstructural damage as well as on the overall deformational and failure response of composites.

Thesis (Ph.D.) - University of Glasgow, 2008. / Ph.D. thesis submitted to the Department of Mathematics, Faculty of Information and Mathematical Sciences, University of Glasgow, 2008. Includes bibliographical references. Print version also available.

<div>Thin walled structures find wide applications in automotive industry as energy absorption devices. A great deal of research has been conducted to design thin walled structures, where the main objective is to reduce peak crushing forces and increase energy absorption capacity. With the advancement of computers and mathematics, it has been possible to develop 2D patterns which when folded turn into complex 3D structures. This technology can be used to develop patterns for getting structures with desired properties. </div><div>In this study, square origami tubes with folding pattern (Yoshimura pattern) is designed and studied extensively using numerical analysis. An accurate Finite Element Model (FEM) is developed to conduct the numerical analysis. A parametric study was conducted to study the influence of geometric parameters on the mechanical properties like peak crushing force, mean crushing force, load uniformity and maximum intrusion, when subjected to dynamic loading. </div><div>The results from this analysis are studied and various conclusions are drawn. It is found that, when the tube is folded with the pattern having specific dimension, the performance is enhanced significantly, with predictable and stable collapse. It is also found that the stiffness of the module varies with geometrical parameters. With a proper study it is possible to develop origami structures with functionally graded stiffness, the performance of which can be tuned as per requirement, hence, showing promising capabilities as an energy absorption device where progressive collapse from near to end impact end is desired.</div><div><br></div>

A non-linear finite element program to simulate the behaviour of infilled frames and plane stress reinforced concrete members under the action of monotonic and cyclic loading has been developed. Steel is modelled as a strain hardening plastic material, and in the concrete model cracking, yielding and crushing are considered. The separation, sliding, and opening and closing of initial gaps at the interfaces between the frame and the infill panels are accounted for by adjusting the properties of interface elements. The non-linear equations of equilibrium are solved using an incremental-iterative technique performed under load or displacement control. The iterative techniques use the standard and modified Newton-Raphson method or the secant Newton method. An automatic load incrementation scheme, line searches, and restart facilities are included. The capabilities of the program have been examined and demonstrated by analysing five reinforced concrete panels, a deep beam, a shear wall, and eight infilled frames. The accuracy of the analytical results was assessed by comparing them with the experimental results and those obtained analytically by other workers and shown to be good. A study of the effects of some material and numerical parameters on the results of analyses of reinforced concrete deep beam has been carried out. Two techniques have been proposed and used to overcome numerical problems associated with local strain concentrations which occur with the displacement control, when path dependent incremental iterative procedures are used for inelastic materials. The displacement control provided with these modifications has been shown to be more efficient than the load control.

This thesis describes a non-linear finite element model suitable for the analysis of reinforced concrete, or steel, structures under general three-dimensional states of loading. The 20 noded isoparametric brick element has been used to model the concrete and reinforcing bars are idealised as axial members embedded within the concrete elements. The compressive behaviour of concrete is simulated by an elasto-plastic work hardening model followed by a perfectly plastic plateau which is terminated at the onset the . crushing. In tension, a smeared crack model with fixed orthogonal cracks has been used with the inclusion of models for the retained post-cracking stress and the reduced shear modulus. The non-linear equations of equilibrium have been solved using an incremental-iterative technique operating under load control. The solution algorithms used are the standard and the modified Newton-Raphson methods. Line searches have been implemented to accelerate convergence. The numerical integration has been generally carried out using 15 point Gaussian type rules. Results of a study to investigate the performance of these rules show that the 15 point rules are accurate and computationally efficient compared with the 27(3X3X3) point Gaussian rule. The three- dimensional finite element model has been used to investigate the problem of elasto-plastic torsion of homogeneous members. The accuracy of the finite element solutions obtained for beams of different cross-sections subjected to pure and warping torsion have been assessed by comparing them with the available exact or approximate analytical solutions. Because the present work is devoted towards the analysis of reinforced concrete members which fail in shear or torsional modes, the computer program incorporates three models to account for the degradation in the compressive strength of concrete due to presence of tensile straining of transverse reinforcement. The numerical solutions obtained for reinforced concrete panels under pure shear and beams in torsion and combined torsion and bending reveal that the inclusion of a model for reducing the compressive strength of cracked concrete can significantly improve the correlation of the predicted post-cracking stiffness and the computed ultimate loads with the experimental results. Parametric studies to investigate the effects of some important material and solution parameters have been carried out. It is concluded that in the presence of a compression strength reduction model, the tension-stiffening parameters required for reinforced concrete members under torsion should be similar to those used for members in which bending dominates.

In this thesis, we develop the a new immersed finite element(IFE) space formed by piecewise linear polynomials defined on sub-elements cut by the actual interface curve for solving elliptic interface problems on interface independent meshes. A group of geometric identities and estimates on interface elements are derived. Based on these geometric identities and estimates, we establish a multi-point Taylor expansion of the true solutions and show the estimates for the second order terms in the expansion. Then, we construct the local IFE spaces by imposing the weak jump conditions and nodal value conditions on the piecewise polynomials. The unisolvence of the IFE shape functions is proven by the invertibility of the well-known Sherman-Morrison system. Furthermore we derive a group of fundamental identities about the IFE shape functions, which show that the two polynomial components in an IFE shape function are highly related. Finally we employ these fundamental identities and the multi-point Taylor expansion to derive the estimates for IFE interpolation errors in L2 and semi-H1 norms. / Master of Science

A nonlinear finite element model of the hot and cold rolling processes has been developed for flat rolling stock with rectangular cross section. This model can be used to analyze the flat rolling of cold and hot steel rectangular strips under a series of different parameters, providing the rolling designer with a tool that he can use to understand the behavior of the steel as it flows through the different passes. The models developed, take into account all of the non-linearities present in the rolling problem: material, geometric, boundary, and heat transfer. A coupled thermal-mechanical analysis approach is used to account for the coupling between the mechanical and thermal phenomena resulting from the pressure-dependent thermal contact resistance between the steel slab and the steel rolls. The model predicts the equivalent stress, equivalent plastic strain, maximum strain rate, equivalent total strain, slab temperature increase, increase in roll temperature, strip length increase, slab thickness % reduction (draft), and strip's velocity increase, for both the cold and hot rolling processes. The FE model results are an improvement over the results obtained through the classical theory of rolling. The model also demonstrates the role that contact, plastic heat generation and friction generated heat plays in the rolling process. The analysis performed shows that the steel in cold rolling can be accurately modeled using the elastic-plastic (solid Prandtl-Reuss) formulation, with a von Mises yield surface, the Praguer kinematic hardening rule, and the Ramberg-Osgood hardening material model. The FE models also demonstrate that the steel in hot rolling can be modeled using the rigid-viscoplastic (flow Levy-Mises) formulation, with a von Mises yield surface, and Shida's material model for high temperature steel where the flow stress is a function of the strain, strain rate, and the temperature. Other important contributions of this work are the demonstration that in cold rolling, plane sections do not remain plane as the classic theory of rolling assumes. As a consequence, the actual displacements, velocity, and stress distributions in the workpiece are compared to and shown to be an improvement over the distributions derived from the classical theory. Finally, the stress distribution in the rolls during the cold rolling process is found, and shown to be analogous to the stress distribution of the Hertz contact problem. / Master of Science

In this thesis an investigation of the shear stiffness and capacity of joints between pre- fabricated concrete elements regarding to different material properties is reported. Two different models of shear key joints, connected to prefabricated walls, were cre- ated in the non-linear finite element software, ATENA 3D, with the aim to estimate a realistic behaviour of the joints regarding to the external loads.

Birds and bats employ a variety of advanced wing motions in the efficient production of thrust. The purpose of this thesis is to quantify the benefit of these advanced wing motions, determine the optimal theoretical wing kinematics for a given flight condition, and to develop a methodology for applying the results in the optimal design of flapping-wing aircraft (ornithopters). To this end, a medium-fidelity, combined aero-structural model has been developed that is capable of simulating the advanced kinematics seen in bird flight, as well as the highly non-linear structural deformations typical of high-aspect ratio wings. Five unique methods of thrust production observed in natural species have been isolated, quantified and thoroughly investigated for their dependence on Reynolds number, airfoil selection, frequency, amplitude and relative phasing. A gradient-based optimization algorithm has been employed to determined the wing kinematics that result in the minimum required power for a generalized aircraft or species in any given flight condition. In addition to the theoretical work, with the help of an extended team, the methodology was applied to the design and construction of the world's first successful human-powered ornithopter. The Snowbird Human-Powered Ornithopter, is used as an example aircraft to show how additional design constraints can pose limits on the optimal kinematics. The results show significant trends that give insight into the kinematic operation of natural species. The general result is that additional complexity, whether it be larger twisting deformations or advanced wing-folding mechanisms, allows for the possibility of more efficient flight. At its theoretical optimum, the efficiency of flapping-wings exceeds that of current rotors and propellers, although these efficiencies are quite difficult to achieve in practice.