<p>An advanced bulk constitutive model is used with a new cohesive zone model that is stress state and rate-dependent in order to simulate the ductile to brittle failure transition in polycarbonate. The cohesive zone model is motivated by experimental evidence that two different critical energies per unit area of crack growth exist in glassy polymers. A higher energy state is associated with ductile failure (slow crack growth), while a lower energy state is associated with brittle failure (fast crack growth). The model is formulated so that as rate or stress state changes within a simulation, the fracture energy and thus fracture mode may also change appropriately. The ductile to brittle transition occurs when the cohesive opening rate is over a threshold opening rate and when the stress state is close to plane strain in a fracture specimen. These effects are coupled. The principal contribution of this work is that this is the first time a single set of material input parameters can predict the transition from slow to fast crack growth as test loading rate and sample thickness are varied. This result enlisted the use of an advanced constitutive model and the new cohesive zone model with rate and stress-state dependencies in three-dimensional finite element analysis.</p> / Dissertation

During foil rolling, back-up and work rolls undergo elastic deformation resulted from the rolling reaction force, which results in non-uniform thickness distribution in the width direction, even causes waves and fracture in the rolled foils. This paper aims to propose a mathematical model for a four-high mill to analyze the elastic deformation of the rolls and discuss the relationship between axial defection of the back-up and work rolls and the rolling conditions, from which the thickness distribution of the product is then predicted. The finite element simulation is also used to analyze the rolling force and roll¡¦s elastic deformation of a four-high mill. From the predicted foil shape, the roll profiles are designed. The mathematical model is validated by comparing the analytical thickness distribution with experiment values. Rolling pass schedules are also designed. From the arrangerement of reductions and heat treatment, experimental results of stainless steels foils with 80£gm thick and 2£gm variation, pure copper foils with 20£gm thick and 2£gm variation, and aluminum foils with 15£gm thick and 3£gm variation are successfully obtained. A rolling technology for foil rolling is developed.

In last decade, white light emitting diodes(LEDs) have become used widely from traditional indicator to general illumination. The increase of its power is the key improving issue. The current light efficiency of white LED about 30%. In other words ,more than 70% of the input electrical energy will be generated in the form of heat. So, how to get rid of the heat damage in high power LED is a severe problem. The finite element analysis is employed to simulate high power white LEDs temperature distribution and thermal stress distributions caused by the dissipated heat. The effects of package parameters, i.e. die attach, solder material, solder thickness, and chip substrate, on the temperature and thermal stress distributions on high power LED packages are simulated and studied in this thesis. A comparison between the 40mil single chip package and the chip on board(CoB) package has also been executed in this study. Simulated results indicate that the highest power of a single 40mil chip package is 7watt. The thermal stress distribution , i.e. the peak value of local thermal stress is over its yield strength, is occurred as the power up to 7watt. Numerical results also reveal that the appropriate fin design can improve the heat dissipation significantly in high power LED package.

This study investigates analytically and experimentally extrusion processes of magnesium hollow tubes by a single-cylinder extrusion machine and double-cylinder extrusion machine. The first part of this study is to conduct analysis and experiment of hollow helical tube extrusion by single-cylinder extrusion machine. Firstly, a design criterion is proposed to determine the forming parameters and discuss the effects of product size, extrusion ratio, billet length, etc. on the mandrel surface stress. The effects of the die bearing part length, angle of rotation, extrusion speed, initial temperature, petal number, etc. on the radial filling ratio are also investigated. Better parameters are chosen from analytical results to conduct hot extrusion experiments for obtaining sound products. Microstructure observation and hardness test are conducted at the cross-section of the product. The experimental values of extrusion load and product¡¦s dimensions are compared with the analytical values to verify the validity of the analytical models. The second part of this study is to conduct analysis and experiment of hollow tubes extrusion by a double-cylinder extrusion machine. The effects of extrusion ratio, billet length, mandrel diameter, etc. on the drawing force on the mandrel and critical conditions without mandrel fracture are discussed.

In recent years, traditional piezoelectric materials have been pushed to the limit in terms of performance because of countless novel applications. This has caused an increased interest in piezoelectric composites, which combine two or more constituent materials in order to create a material system that incorporates favorable attributes from each constituent. One or more of the constituents exhibits piezoelectric behavior, so that the composite has an effective electromechanical coupling. The composite material may also have enhanced properties such as stiffness, durability, and flexibility. Finite element analyses were conducted on a three-phase piezoelectric nanocomposite in order to investigate the effects of several design parameters on performance. The nanocomposite consisted of a polyimide matrix, beta-CN APB/ODPA, enhanced with single wall carbon nanotubes and PZT-5A particles. The polyimide and nan- otube phases were modeled as a single homogenized phase. This results in a two-phase nanocomposite that can be modeled entirely in the continuum domain. The material properties for the nano-reinforced matrix and PZT-5A were obtained from previous experimental efforts and from the literature. The finite element model consisted of a single representative volume element of the two-phase nanocomposite. Exact periodic boundary conditions were derived and used to minimize the analysis region. The effective mechanical, electrical, and piezoelectric properties were computed for a wide range of nanotube and PZT particle concentrations. A discrepancy was found between the experimental results from the literature and the computational results for the effective electrical properties. Several modified finite element models were developed to explore possible reasons for this discrepancy, and a hypothesis involving dispersion of the nanotubes was formulated as an attempt to explain the difference. The response of the nanocomposite under harmonic loading was also investigated using the finite element model. The effective properties were found to be highly dependent on the dielectric loss of the beta CN/SWNT matrix. It was also found that increasing the matrix loss enhanced piezoelectric performance up to a certain point. Exploiting this type of behavior could be an effective tool in designing piezoelectric composite materials.

The tube hydro-forming (THF) is an innovative manufacturing process which is used to manufacture many industrial components widely. The success of THF is largely dependent on the selection of the loading paths: internal pressure versus time, axial feeding versus time and counter punch (CP) versus time. The finite element analysis is used to simulate the forming result of different loading paths and reduce the cost of die-testing. This paper presents the forming of T-branches and T-branches components with CP. These paper has developed an adaptive simulation algorithm by combining FEM code LS-DYNA 3D with controller subroutine to get ideal bulging height and uniform thickness of the formed tube with multi-stages. Discuss influence under different parameters of process. The results are compared with experimental results to validate accuracy by this adaptive control methods.

CPU heat sinks with high efficiency of heat transfer are greatly demanded for a personal computer with high-speed computational ability. In recent years, the manufacturing technology of CPU heat sinks has got much attention and becomes indispensable for developing the high-performance CPUs.In this study, some different design criteria for the flow guide and die are proposed during an extrusion process with complex cross-sectional shapes, such as CPU heat sinks. The plastic flow pattern of the billet inside the die cavity is analyzed by using a commercial finite element package ¡§DEFORM 3D¡¨.The extrusion load, the stress and strain distribution of die, and the curvature of the product are investigated. Taguchi method is used to find the optimum extrusion condition of the die parameters. In addition, the experiments of extrusion using Al 6061 were carried out. The plastic flow pattern of the billet within the die and the dead metal zones were observed. The experimental data were compared with the analytical values to verify the validity of the proposed analytical models.

The present study aims to determine the optimum blank shape design for the deep drawing of arbitrary shaped cups with a uniform trimming allowance at the flange i.e. cups without ears. This earing defect is caused by planar anisotropy in the sheet and the friction between the blank and punch/die. In this research, a new method for optimum blank shape design using finite element analysis has been proposed. Explicit non-linear finite element (FE) code LSDYNA is used to simulate the deep drawing process. FE models are constructed incorporating the exact physical conditions of the process such as tooling design like die profile radius, punch corner radius, etc., material used, coefficient of friction, punch speed and blank holder force. The material used for the analysis is mild steel. A quantitative error metric called shape error is defined to measure the amount of earing and to compare the deformed shape and target shape set for each stage of the analysis. This error metric is then used to decide whether the blank needs to be modified or not. The cycle is repeated until the converged results are achieved. This iterative design process leads to optimal blank shape. In order to verify the proposed method, examples of square cup and cylindrical cup have been investigated. In every case converged results are achieved after a few iterations. So through the investigation the proposed systematic method of optimal blank design is found to be very effective in the deep drawing process and can be further applied to other stamping applications.

In the present study, we propose a computational model for the linear and nonlinear analysis of shell structures. We consider a tensor-based finite element formulation which describes the mathematical shell model in a natural and simple way by using curvilinear coordinates. To avoid membrane and shear locking we develop a family of high-order elements with Lagrangian interpolations. The approach is first applied to linear deformations based on a novel and consistent third-order shear deformation shell theory for bending of composite shells. No simplification other than the assumption of linear elastic material is made in the computation of stress resultants and material stiffness coefficients. They are integrated numerically without any approximation in the shifter. Therefore, the formulation is valid for thin and thick shells. A conforming high-order element was derived with 0 C continuity across the element boundaries. Next, we extend the formulation for the geometrically nonlinear analysis of multilayered composites and functionally graded shells. Again, Lagrangian elements with high-order interpolation polynomials are employed. The flexibility of these elements mitigates any locking problems. A first-order shell theory with seven parameters is derived with exact nonlinear deformations and under the framework of the Lagrangian description. This approach takes into account thickness changes and, therefore, 3D constitutive equations are utilized. Finally, extensive numerical simulations and comparisons of the present results with those found in the literature for typical benchmark problems involving isotropic and laminated composites, as well as functionally graded shells, are found to be excellent and show the validity of the developed finite element model. Moreover, the simplicity of this approach makes it attractive for future applications in different topics of research, such as contact mechanics, damage propagation and viscoelastic behavior of shells.

Research is performed to assess the potential of surface moisture evaporative cooling from composite plates as a means of reducing the external temperature of military aircraft. To assess the feasibility of evaporative cooling for this application, a simplified theoretical model of the phenomenon is formulated. The model consists of a flat composite plate at an initial uniform temperature, T0. The plate also possesses an initial moisture (molecular water) content, M0. The plate is oriented vertically and at t=0 s, one surface is exposed to a free stream of air at an elevated temperature. The other surface is exposed to stagnant air at the same temperature as the plate’s initial temperature. The equations associated with energy and mass transport for the model are developed from the conservation laws per the continuum mechanics hypothesis. Constitutive equations and assumptions are introduced to express the two nonlinear partial differential equations in terms of the temperature, T, and the partial density of molecular water, &#961;w. These equations are approximated using a weak form Galerkin finite element formulation and the &#945;–family of time approximation. An algorithm and accompanying computer program written in the Matlab programming language are presented for solving the nonlinear algebraic equations at successive time steps. The Matlab program is used to generate results for plates possessing a variety of initial moisture concentrations, M0, and diffusion coefficients, D. Surface temperature profiles, over time, of moisture bearing specimens are compared with the temperature profiles of dry composite plates. It is evident from the results that M0 and D affect the surface temperature of a moist plate. Surface temperature profiles are shown to decrease with increasing M0 and/or D. In particular, dry and moist specimens are shown to differ in final temperatures by as much as 30°C over a 900 s interval when M0 = 30% and D is on the order of 10–8m2/s (T0 = 25°C, h = 60 W/m2°C, T&#8734; = 90°C).