The aim of the project presented in this thesis is to demonstrate a modelling method for predicting the variability in the ultimate load of stiffened panel under axial compression due to manufacturing variability. Bulking is sensitive to imperfections. In the case of a post-buckled panel, manu-facturing variability produces a scatter in the ultimate load. Thus, reasonable leeway for imperfections and inherent variability must be allowed in their design. Firstly, a finite element model of a particular stiffened panel was developed, and all nonlinearities within the material, boundary condition and geometry were considered. Verification and validation were performed to examine the accuracy of the buckling behaviour prediction, especially ultimate load. Experiments on 5 identical panels in design were performed to determine the level of panel-panel variation in geometry and collapse load. A data reduction programme based on the practical geometry scanning was developed, in addi-tion to which, the procedure of importing measured imperfection into Finite Ele-ment model was introduced. To identify and apply representative imperfections to the panel model, a double Fourier series representation of the random geometric distributions is attempt-ed, and was used thereby to derive a series of shapes representing random ge-ometry scatters. With these newly generated geometric imperfections, the variation in collapse load was determined, using the validated FE analysis. And also, the probability of these predicted loads was generalized.

The analysis and design of bracing systems for complex frame geometries typically found in metal buildings can prove to be an arduous task given current methods. The American Institute of Steel Construction's Appendix 6 from the 2010 Specification for Structural Steel Buildings affords engineers a means for determining brace strength and stiffness requirements, but only for the most basic cases. Specifically, there are a number of aspects of metal building systems that place their designs outside the scope of AISC's Appendix 6 (Stability Bracing for Columns and Beams). Some of the aspects not considered by Appendix 6 include: the use of web-tapered members, the potential for unequally spaced or unequal stiffness bracing, combination of bracing types including panel and flange diagonal bracing, and the effects of continuity across brace points. In this research, an inelastic eigenvalue buckling procedure is developed for calculation of the ideal bracing stiffness demands in general framing systems. Additionally, the software provides a method of calculating the elastic lateral-torsional buckling load of members with generally stepped and tapered cross-sections, which satisfies an important need for rigorous design assessment. Extensive benchmarking to load-deflection simulations of geometrically imperfect systems is performed and recommendations are developed for determining the required design stiffness and strength of the bracing components based on the use of this type of computational tool.

The purpose of this thesis work is to evaluate how the physical behaviour of a pipe bend is affected by the pipe bending procedure. Effects such as initial ovalization, thinning, thickening and plastic hardening from the bending procedure are examined and the mechanical properties of pipe bends containing these effects are investigated. This has been evaluated by creating a detailed Finite Element model of a pipe that is being bent. Then the differences compared to a bent tube in a virgin state, so called Elbow elements and an analytical in-house program have been evaluated. The virgin state refers to a model of a pipe that is bent from the beginning, thus having a homogeneous thickness and not containing any plastic hardening. The Elbow element is a calculationally cheap element, specially developed for accurate calculations of pipe bends in an initially virgin state. The goal with the thesis work is to get a better picture of what happens to a pipe as it is being bent, how this affects the mechanical properties and to evaluate the possibility to develop an easy method for taking these effects into account when using the Elbow element. This report describes the layout of the work and how the detailed FE-model has been constructed. One step to being able to use the Elbow element with respect to changes in shape and plastic hardening from the manufacturing process has been presented, the differences are though considered being too big to be able to use the Elbow elements with enough confidence in the results. The problems that remain are presented and discussed and proposals for further work are presented.

This Master’s thesis is summery of a comparative study of 2 commercial CAD-integrated FE-programs. These FE-programs were CATIA v5 and ABAQUS 6.3-7. The primary objective of this study is to investigate the basic FEA capabilities of CATIA and ABAQUS 6.7-3 in performing the linear static analysis and to identify whether there are any differences and similarities between results the both Finite Element FE codes give. The overall research question in the present thesis is: Do different FE programs, here CATIA and ABAQUS, give the same results for FE analysis giving the same models if subjected to the same boundary conditions? This research seeks to achieve its aims through making a comparative qualitative study. Certain pre-selections were performed in advance of conducting Finite element analysis and the comparison process to ensure that results would reflect only the most relevant and meaningful differences and similarities between the both FE-codes. Five different 3D solid models have been selected to perform linear static Finite element analysis on. All these models (case studies) are created in CATIA V5 and the linear static analysis conducted on using FE-codes CATIA v5 and ABAQUS 6.7-3. Three static responses (results) of the linear static analysis have been adopted as criteria for comparisons purposes. These criteria were: (1) displacements, (2) Von Mises stress, and (3) principal stress. The results of comparisons showed that there is a very good agreement in most cases and small gap between in a few cases. Results of this study demonstrate that the both FE-programs CATIA v5 and ABAQUS 6.7-3 have good capabilities to perform FE-analysis and they give very near results. Reason behind differences is that each of them uses a different algorithm for solving problems. The final answer for the research question is given with valuable recommendations for future work in the scope of this research.

This work deals with the bearing capacity of wood based I-joists Finite element models were analyzed to determine the bearing capacity of I-joists, using the finite element software Abaqus CAE. The purpose of this study is to compare the results from the developed FE-models with experimental results, and with a previously proposed design formula. To perform the analyses finite element models were created. The model consists of three parts:, the web (made of shell element), the flanges and steel plates used at the supports and loading points (made of solid elements) To determine the bearing capacity of the I-joist two types of analyses were performed, a linear buckling analysis to check the risk of web buckling and a static (stress) analysis to check the risk of splitting of the flanges. This study shows that the steel plate length, in some cases, has little or no impact on primarily the splitting load. Furthermore, the buckling load decreases as the depth of the beam increases, the influence of the depth being proportional to 1/h2. The depth of the beam has no impact on the risk of splitting of the flange.

The goal in this thesis was to reduce a multilayer shim model, which was modeled from steel and polymer (isotropic materials), into an equivalent single layer shim model. The procedure was to use mathematical formulations to convert a multilayer shim into an ESL (equivalent single layer) shim. Here, a transverse isotropic model is used to prepare for future orthotropic layers. The results show that the ESL model behaves isotropically. In the 2 layer model there was no squeal noise whereas in the ESL models there is.

In this study, the settlement behavior of the piled raft foundations resting on overconsolidated clays under uniform loading, is investigated for different pile configurations and load levels. A total of 100 plane &ndash / strain and three &ndash / dimensional finite element analyses are carried out and the results of these analyses are compared both with each other and with the results presented by Reul &amp / Randolph (2004). The material parameters used in the analysis are selected mainly referring to the previous studies cited above on the same subject and slight modifications are made for convenience in the analysis. The analysis method and the applied pile configurations and load levels are directly taken from the reference study, excluding the soil model employed. A drained Mohr &ndash / Coulomb failure criteria is employed in the analysis of this study in modeling the soil instead of an elastoplastic model which was used in the analysis of the reference study. The results are evaluated for the average and differential settlements of the foundations and it is seen that / although the average and differential settlements calculated in this study are not always very close to the values calculated in the reference study, the calculated settlement reduction factors due to piles (especially for the average settlements) compared well with the findings of the reference study for all pile configurations and load levels considered. Based on this, a new approach is suggested to estimate the average settlements of the piled raft foundations. Moreover, correction factors are recommended in order to estimate the average settlements of the piled rafts by directly using the programs employed throughout the thesis.

<p>In-service inspections of high temperature pressureequipment show that weldments are prone to creep and fatiguedamage. It is not uncommon that severely damaged weldments arefound even before the design life of the component has beenreached. In order to improve this situation action has beentaken during the last decades, both from industry, universitiesand research institutes, aiming at an enhanced understanding ofthe weldment response.</p><p>The work presented in this thesis focuses on numericalsimulation of weldment creep response. For a more profoundunderstanding of the evolution of creep damage in mismatchedlow alloy weldments, simulations are performed using thecontinuum damage mechanics, CDM, concept. Both design and lifeassessment aspects are addressed. The possibility to assessseam welded pipes using results from tests of cross-weldspecimens taken out from the seam is investigated. It is foundthat the larger the cross-weld specimen the better thecorrelation. The advantage to use the CDM concept prior to aregular creep analysis is also pointed out. In order to developthe CDM analysis, a modified Kachanov-Rabotnov constitutivemodel is implemented into ABAQUS. Using this model, a secondredistribution of stresses is revealed as the tertiary creepstage is reached in the mismatched weldment.</p><p>Creep crack growth, CCG, in cross-weld compact tension, CT,specimens is investigated numerically where a fracturemechanics concept is developed in two steps. In the first one,the C^*value and an averaged constraint parameter areused for characterising the fields in the process zone, whilein the second step, the creep deformation rate perpendicular tothe crack plane and a constraint parameter ahead of the cracktip, are used as characterising parameters. The influence oftype and degree of mismatch, location of starter notch as wellas size of CT specimen, is investigated. Results show that notonly the material properties of the weldment constituentcontaining the crack, but also the deformation properties ofthe adjacent constituents, influence the CCG behaviour.Furthermore, the effect of size is influenced by the mismatchof the weldment constituents.</p><p>A circumferentially cracked girth weld with differentmismatch is assessed numerically by use of the fracturemechanics concept developed. The results show that type anddegree of mismatch have a great influence on the CCG behaviourand that C^*alone cannot characterise crack tip fields.Corresponding R5 assessments are also performed. Comparisonwith the numerical investigation shows that the assumption ofplane stress or plane strain conditions in the R5 analysis isessential for the agreement of the results. Assuming the formerresults in a relatively good agreement for the axial stressdominated cases while for the hoop stress dominated cases, R5predicts higher CCG rates by an order of magnitude.</p><p><b>Keywords:</b>ABAQUS, constraint effect, continuum damagemechanics, creep, creep crack growth, design, design code,finite element method, fracture mechanics, life assessment,mismatch, numerical simulation, weldment</p>

<p>The purpose of this thesis work is to evaluate how the physical behaviour of a pipe bend is affected by the pipe bending procedure. Effects such as initial ovalization, thinning, thickening and plastic hardening from the bending procedure are examined and the mechanical properties of pipe bends containing these effects are investigated.</p><p>This has been evaluated by creating a detailed Finite Element model of a pipe that is being bent. Then the differences compared to a bent tube in a virgin state, so called Elbow elements and an analytical in-house program have been evaluated. The virgin state refers to a model of a pipe that is bent from the beginning, thus having a homogeneous thickness and not containing any plastic hardening. The Elbow element is a calculationally cheap element, specially developed for accurate calculations of pipe bends in an initially virgin state.</p><p>The goal with the thesis work is to get a better picture of what happens to a pipe as it is being bent, how this affects the mechanical properties and to evaluate the possibility to develop an easy method for taking these effects into account when using the Elbow element.</p><p>This report describes the layout of the work and how the detailed FE-model has been constructed. One step to being able to use the Elbow element with respect to changes in shape and plastic hardening from the manufacturing process has been presented, the differences are though considered being too big to be able to use the Elbow elements with enough confidence in the results. The problems that remain are presented and discussed and proposals for further work are presented.</p>

Thesis--Universität Paderborn. / Includes bibliographical references.