This paper describes a method by means of FE code DEFORMTM-2D to simulate the formation of central bursting defects in extrusion processes; the effect of various extrusion parameters such as half die angle, reduction in area, friction factor, and strain hardening exponent on the maximum damage value is examined. The differences between various ductile fracture criteria are compared and critical damage value(CDV) of the material AA6061 is found. In addition, we get the strength coefficient(K), strain hardening exponent(n), CDV and friction factor(m) by material tests, such as uniform tensile test, notched tensile test, compression test, and ring compression test. Finally, the cold multistage extrusion experiment was conducted to verify the accuracy of the finite element simulations. From the continuous three pass extrusion experimental data, no fracture in the center of the extruded product was found. From the analytical data, it was known that the maximum damage value 1.0479 for third pass extrusion was small than critical damage value 1.068, thus, central bursting defects didn¡¦t occur in extrusion processes.

Surface-crack configurations are among the most important crack problems in the aerospace industry. The residual strength of a surface-cracked component is complicated by three-dimensional variation of the stress-intensity factor around the crack front and plastic deformations, which vary from plane stress at the free boundary, to nearly plane-strain behavior in the interior. In 1973, a two-parameter fracture criterion (TPFC) was developed to analyze fracture behavior of surface-crack configurations. Estimates were made around the crack front for fracture initiation—the critical parametric angle. Recently, NASA developed the Tool for Analysis of Surface Cracks (TASC) software that predicts critical location. This thesis is the application of the TPFC with the TASC critical angles using an equation developed from the TASC software. The TPFC was applied to three materials: a brittle titanium alloy, a ductile titanium alloy, and a ductile 301 stainless steel. The TPFC with the TASC critical angles correlated fracture behaviors well.

O presente trabalho visa o estudo e avaliação de diversos modelos de falha através de ensaios experimentais e análises numéricas. A caracterização do material foi feita por meio de ensaios de tração de espécimes com e sem entalhe e de cisalhamento, para prover dados em uma ampla faixa de triaxialidade. A calibração dos parâmetros necessários para cada modelo de falha ocorreu através da obtenção das componentes de tensões e de deformações na região de ruptura, obtidas por meio de análises numéricas destes ensaios. O critério da máxima deformação plástica equivalente, modelo de dano de Johnson-Cook e critério da máxima tensão cisalhante foram aplicados em um caso de impacto em pára-choque. Diversos ensaios de impacto foram realizados até a que se atingisse a ruptura satisfatória do componente. Os ensaios foram modelados em elementos finitos, através do programa comercial LS-Dyna®, sendo que os modelos de falha calibrados foram aplicados para o ensaio onde se ocorreu ruptura. Houve boa concordância entre os resultados obtidos numericamente e experimentalmente, respeitadas as observações realizadas acerca da especificidade deste caso de impacto em baixa velocidade. / The aim of the present work is an assessment of several fracture models through experiments and numerical simulations. Tensile tests with notched and unnotched specimens were carried out to provide the material characterization in a wide range of stress triaxiality. The calibration of the parameters required by the fracture models was enabled by the parallel numerical simulation of the tensile tests, providing information on the stress and strain components at the failure locus. The constant equivalent strain criterion, the Johnson-Cook failure model and the maximum shear stress failure criterion were applied in a bumper beam impact case study. Several low speed impact tests were carried out in order to result in the components rupture. Numerical simulation of the experiments was performed using commercial finite element code LS-Dyna®. Good correlation of experiments and numerical simulations was achieved when considering this particular low speed case study.

O presente trabalho visa o estudo e avaliação de diversos modelos de falha através de ensaios experimentais e análises numéricas. A caracterização do material foi feita por meio de ensaios de tração de espécimes com e sem entalhe e de cisalhamento, para prover dados em uma ampla faixa de triaxialidade. A calibração dos parâmetros necessários para cada modelo de falha ocorreu através da obtenção das componentes de tensões e de deformações na região de ruptura, obtidas por meio de análises numéricas destes ensaios. O critério da máxima deformação plástica equivalente, modelo de dano de Johnson-Cook e critério da máxima tensão cisalhante foram aplicados em um caso de impacto em pára-choque. Diversos ensaios de impacto foram realizados até a que se atingisse a ruptura satisfatória do componente. Os ensaios foram modelados em elementos finitos, através do programa comercial LS-Dyna®, sendo que os modelos de falha calibrados foram aplicados para o ensaio onde se ocorreu ruptura. Houve boa concordância entre os resultados obtidos numericamente e experimentalmente, respeitadas as observações realizadas acerca da especificidade deste caso de impacto em baixa velocidade. / The aim of the present work is an assessment of several fracture models through experiments and numerical simulations. Tensile tests with notched and unnotched specimens were carried out to provide the material characterization in a wide range of stress triaxiality. The calibration of the parameters required by the fracture models was enabled by the parallel numerical simulation of the tensile tests, providing information on the stress and strain components at the failure locus. The constant equivalent strain criterion, the Johnson-Cook failure model and the maximum shear stress failure criterion were applied in a bumper beam impact case study. Several low speed impact tests were carried out in order to result in the components rupture. Numerical simulation of the experiments was performed using commercial finite element code LS-Dyna®. Good correlation of experiments and numerical simulations was achieved when considering this particular low speed case study.

A two-parameter fracture criterion (TPFC) is used to correlate and predict failure loads on cracked configurations made of ductile materials. The current study was conducted to validate the use of the fracture criterion on more brittle materials, using elastic-plastic finite-element analyses with the critical crack-tip-opening angle (CTOA) failure criterion. Forman generated fracture data on middle-crack tension, M(T), specimens made of thin-sheet 7075-T6 aluminum alloy, which is a quasi-brittle material. The fracture data included a wide range of specimen widths (2w) ranging from 3 to 24 inches. A two-dimensional (2D) finite-element analysis code (ZIP2D) with a ''plane-strain core" option was used to model the fracture process. Fracture simulations were conducted on M(T), single-edge-crack tension, SE(T), and single-edge-crack bend, SE(B), specimens. The results supported the TPFC equation for net-section stresses less than the material proportional limit. However, some discrepancies were observed among the numerical results of the three specimen types. Thus, more research is needed to improve the transferability of the TPFC from the M(T) specimen to both the SE(T) and SE(B) specimens.

The Two-Parameter-Fracture-Criterion (TPFC) was validated using an elastic-plastic two-dimensional (2D) finite-element code, ZIP2D, with the plane-strain-core concept. Fracture simulations were performed on three crack configurations: (1) middle-crack-tension, M(T), (2) single-edge-crack-tension, SE(T), and (3) single-edge crack-bend, SE(B), specimens. They were made of 2014-T6 (TL) aluminum alloy. Fracture test data from Thomas Orange work (NASA) were only available on M(T) specimens (one-half width, w = 1.5 to 6 in.) and they were all tested at cryogenic (-320oF) temperature. All crack configurations were analysed over a very wide range of widths (w = 0.75 to 24 in.) and crack-length-to-width ratios ranged from 0.2 to 0.8. The TPFC was shown to fit the simulated fracture data fairly well (within 6.5%) for all crack configurations for net-section stresses less than the material proportional limit. For M(T) specimens, a simple approximation was shown to work well for net-section stresses greater than the proportional limit. Further study is needed for net-section stresses greater than the proportional limit for the SE(T) and SE(B) specimens.

Dans le cas du bois massif, les résistances en flexion et en traction peuvent être altérées par la présence de défauts. Les nœuds et la distorsion du fil qui en découle, sont les causes de la diminution du rendement mécanique des sections de poutres de structure. A ce jour, l’effet de ces défauts sur la perte de résistance de l’élément de structure ne peut être quantifié avec exactitude. Le classement mécanique du bois de structure n’est alors pas optimal et ne permet pas de valoriser les éléments de fortes résistances. La présente étude a pour objectif d’étudier l’influence de la nodosité sur le rendement mécanique de poutres en Pin Maritime dans le but d’en améliorer son classement mécanique. Cette étude s’appuie sur une base de données recensant les propriétés mécaniques et physiques de poutres de grandes dimensions ainsi que les propriétés géométriques des nœuds présents sur ces éléments. Différents outils statistiques sont utilisés afin de définir les variables explicatives et leurs pouvoirs prédictifs sur la résistance mécanique de ces composants structurels. De ces analyses, il ressort que certaines caractéristiques propres aux nœuds sont significatives et permettent d’améliorer la prédiction de la résistance. Des modélisations éléments finis de tronçons de poutres présentant un nœud sont proposées pour justifier la présence de ces paramètres dans les modèles prédictifs et afin de compléter cette base de données par des propriétés mécaniques. Ces modélisations numériques, accompagnées d’un critère de ruine, permettent d’établir des courbes de performances mécaniques en présence d’un nœud. Ces données numériques sont comparées aux résultats expérimentaux afin de vérifier la fiabilité de l’approche proposée. / The bending and tensile strengths of timber can be reduced by the presence of heterogeneities. The knots and the local slope of grain in the surrounding area are the most relevant. Nowadays, the effect of knots reduces the stiffness and the strength of structural elements. The Timber grading is not optimized and does not promote the strongest beams for a timber engineering use. The goal of this study is to evaluate the influence of knots on the bending strength of beams in Maritime Pine, order to improve their timber grading. This work is focused on a database, which takes into account the mechanical and the physical properties of beams and specific geometrical parameters relative to knots detected in elements. Several statistical tools like neural networks are used to define the explanatory variables and their predictive powers of the bending strength. From these analyses, it is shown that some variables characterizing the knots are pertinent and allow the improvement of the prediction of the strength. To justify the presence of these variables in the predictive equations, finite element simulations of single knot are proposed. These numerical computations associated to a fracture criterion allow the establishment of strength performance curves due to the presence of knot in an elementary loaded volume (tension and bending configurations). These numerical data are compared to the experimental results in the aim to verify the reliability of the finite element approach.

The thesis deals with computational modelling of ceramic foams and analysis of influence of structural manufacturing defects (like broken struts, closed pores and material clumps) have on foam strength. Model of foam geometry was discretized using beam elements in order to decrease computational cost. In place where several struct join, rigid beam element was used so that the increased stiffness is better modelled. Closed walls of pores were modelled and discretised by shell elements. Influence of loading direction was analysed on foams containing no defects and then influence of amount of defects in foam on strength was further analysed. Highest strength show foams created by cells whose structs are oriented in direction of loading. These were losing strength most rapidly. Foam with structure of rhombic dodecahedral cell was least influenced by presence of closed pore defects but it also showed lowest strength even without defects. Cells with struts oriented in direction of loading experienced biggest drop in strength. Kelvin cell is a compromise. It was shown that difference in strength of strut with constant and varying cross-section is tenths of percent.

The master’s thesis deals with a numerical simulation of failure of ceramic foams with open-cell structure and with understanding of conditions required for the failure of the structure under various mechanical loading conditions. To this purpose, the so-called stress-energy coupled criterion was utilized. The motivation for this thesis was to create a model able of the most accurate prediction of the ceramic foam strength in comparison with experimental observations. First part of the thesis is focused on the theoretical background required for solving the problem. More specifically there are mentioned methods of the foam material modelling, Linear Elastic Fracture Mechanic (LEFM) and coupled stress-energy criterion used for definition of the crack initiation. In the second part of the thesis, numerical Finite Element Analyses (FEA) whose main purpose was to determine critical conditions necessary for the initiation of strut failure within the foam structure, were performed. These pieces of knowledge were then used for creation of the numerical simulation algorithm of the mechanical test of foam material with regular cell pattern. Outputs of numerical simulations were at the end of this work compared with experimental results (of the compression test) made on the real Al_2 O_3 foams prepared by 3D printing technology and provided by the Institute of Physics of Materials Czech Academy of Science. It can be concluded that a good agreement between results of both approaches was reached and the prediction of the ceramic foam mechanical strength using the developed model is in the meanwhile the most accurate estimation from recently published approaches.

The focus of this thesis is to develop a simulation model for predicting ductile fractures during hot working at Alleima. The main fracture mechanism in these conditions is ductile fracture by void coalescence. The ductile fractures are caused by the linking of voids that appear when there is large plastic deformation near second-phase particles. The chosen method to simulate these was to use a Ductile Fracture Criterion (DFC), which builds on using FE models with a damage parameter. Two criteria were selected to be tested. The austenitic stainless-steel alloy 316L was selected as material for this work. Using the Gleeble 3500 system, hot tension and compression experiments were performed to gather data needed for the simulation models as well as inducing ductile fractures. Rupture occurred for all the hot tension samples and cracks were found for only one of the hot compression experiments. Using data from the Gleeble tests, a separate simulation model for each of the setups were created using the finite element software Marc/Mentat. A flow stress model for 316L was developed. Results from the simulations show that both selected DFCs can be used to predict ductile fractures. Particularly for hot tension. It was shown that it is important to model the temperature gradient in the sample accurately. For hot compression, it was difficult to conclude if the criterions were able to predict fracture since only one data point was available. The thesis concludes that there could be of interest with continued work using DFCs at Alleima.