Refine
Document Type
- Conference Proceeding (24) (remove)
Keywords
- Finite-Elemente-Methode (24) (remove)
Wirklichkeitsnahe Erfassung und Beschreibung des Trag- und Verformungsverhaltens von Strukturen baulicher Anlagen hat in den letzten Jahrzehnten ständig an Bedeutung gewonnen. Konstruktionen im Hoch- und Industriebau werden zunehmend multifunktional genutzt - die >Grenzen< zwischen Bauwerk und Tragwerk, zwischen Hüll- und Tragkonstruktion lösen sich auf. Werden raumabschließende Elemente (Wände, Decken, Dächer) gleichzeitig als Tragelemente und wärme- und schalldämmende Konstruktionen ausgeführt, so entstehen beispielsweise Sandwichplatten, deren Schichten sehr stark differierende Materialeigenschaften aufweisen. Beim Aufbau des FEM-Modells für vielschichtige Schalen können die Formänderungshypothesen für jede Schicht einzeln als auch für die Schale insgesamt gegeben werden. Im ersten Fall ist der Knotenfreiheitsgrad von der Schichtenzahl abhängig, im zweiten Fall nicht. Im weiteren wird eine Formänderungshypothese für das Schichtenpaket angenommen. Ausgegangen wird von den Gleichungen der 3D-Elastizitätstheorie. Die Berücksichtigung der Querkraftschubverformungen ergibt die Möglichkeit einer adäquaten Beschreibung der Verformungen sowohl dünner Schalen als auch von Schalen mittlerer Dicke; die Berechnung der Krümmungen und der LAMEschen Parameter der Bezugsfläche zu umgehen, was für komplizierte Schalenformen eine selbständige Aufgabe ist; eines natürlichen Übergangs von homogenen zu geschichteten Schalen. Das vielschichtige isoparametrische Schalen-FE wird vorgestellt, seine Implementierung in das in Entwicklung befindliche Programmsystem SLANG wird vorbereitet.
For the analysis of arbitrary, by Finite Elements discretized shell structures, an efficient numerical simulation strategy with quadratic convergence including geometrically and physically nonlinear effects will be presented. In the beginning, a Finite-Rotation shell theory allowing constant shear deformations across the shell thickness is given in an isoparametric formulation. The assumed-strain concept enables the derivation of a locking-free finite element. The Layered Approach will be applied to ensure a sufficiently precise prediction of the propagation of plastic zones even throughout the shell thickness. The Riks-Wempner-Wessels global iteration scheme will be enhanced by a Line-Search procedure to ensure the tracing of nonlinear deformation paths with rather great load steps even in the post-peak range. The elastic-plastic material model includes isotropic hardening. A new Operator-Split return algorithm ensures considerably exact solution of the initial-value problem even for greater load steps. The combination with consistently linearized constitutive equations ensures quadratic convergence in a close neighbourhood to the exact solution. Finally, several examples will demonstrate accuracy and numerical efficiency of the developed algorithm.
To fulfil safety requirements the changes in the static and/or dynamic behaviour of the structure must be analysed with great care. These changes are often caused by local reduction of the stiffness of the structure caused by the irregularities in the structure, as for example cracks. In simple structures such analysis can be performed directly, by solving equations of motion, but for more complex structures a different approach, usually numerical, must be applied. The problem of crack implementation into the structure behaviour has been studied by many authors who have usually modelled the crack as a massless rotational spring of suitable stiffness placed at the beam at the location where the crack occurs. Recently, the numerical procedure for the computation of the stiffness matrix for a beam element with a single transverse crack has been replaced with the element stiffness matrix written in fully symbolic form. A detailed comparison of the results obtained by using 200 2D finite elements with those obtained with a single cracked beam element has confirmed the usefulness of such element.
Detailuntersuchungen an Tragwerken führen bei FE-Berechnungen immer wieder auf das Problem einer geeigneten Netzgestaltung. Während in weiten Bereichen ein grobes Netz ausreicht, muß an kritischen Stellen ein sehr feines Netz gewählt werden, um gerade dort hinreichend genaue Ergebnisse zu erhalten. Bei der Realisierung lokaler Netzverdichtungen stellt die Gestaltung des Übergangs vom groben zum feinen Netz das Hauptproblem dar. Im Beitrag wird hierzu eine Familie von FE-Übergangselementen vorgestellt, mit denen sich eine voll-kompatible Kopplung von wenigen großen Elementen mit vielen kleinen Elementen bereits über nur eine Stufe erzielen läßt. Diese neu entwickelten sogenannten pNh-Elemente ermöglichen an einer oder mehreren Seiten den Anschluß von N kleineren Elementen (Elementseiten für h-Verfeinerung). Das wird durch N stückweise definierte Ansatzfunktionen an den entsprechenden Seiten erreicht, wobei die Teilung nicht äquidistant sein braucht. Darüber hinaus ist es möglich, Elemente unterschiedlichen Polynomgrades p an den Standardseiten und den Verfeinerungsseiten anzuschließen. Der praktische Einsatz der Übergangselemente setzt geeignete automatische oder halbautomatische Netzgeneratoren voraus, die diese Elemente einbeziehen. Im Rahmen einer substrukturorientierten Modellierung läßt sich dies besonders günstig realisieren. Im Beitrag wird gezeigt, wie durch Zerlegung des Gesamtmodells in Bereiche mit grobem Netz, mit Übergangsnetz und mit feinem Netz, eine effektive Generierung der Netzverdichtungen zu erreichen ist. An einem praktischen Beispiel aus dem Bauingenieurwesen werden die Vorteile des vorgestellten Übergangselementkonzeptes umfassend demonstriert.
The paper deals with the simulation of the non-linear and time dependent behaviour of complex structures in engineering. Such simulations have to provide high accuracy in the prediction of deformations and stability, by taking into account the long term influences of the non-linear behaviour of the material as well as the large deformation and contact conditions. The limiting factors of the computer simulation are the computer run time and the memory requirement during solving large scale problems. To overcome these problems we use a dynamic-explicit time integration procedure for the solution of the semi-discrete equations of motion, which is very suited for parallel processing. In the paper at first we give a brief review of the theoretical background of the mechanical modelling and the dynamic-explicit technique for the solution of the semi-discrete equations of motion. Then the concept of parallel processing will be discussed . A test example concludes the paper.
Discrete-continual Finite Element Method of Analysis for Three-dimensional Curvilinear Structures
(2003)
This paper is devoted to discrete-continual finite element method (DCFEM) of analysis for three-dimensional curvilinear structures. Operational and variational formulations of the problem in the ring coordinate system are presented. The discrete-continual design model for structures with constant physical and geometrical parameters in longitudinal direction is offered on the basis of so-called curvilinear discrete-continual finite elements. Element coordinate system, approximation of nodal unknowns, construction of element nodal load vector are under consideration. Element system of differential equations is formulated with use of special generalized block-structured stiffness matrix of discrete-continual finite element. Local differential relations are formulated. Resultant multipoint boundary problem for system of ordinary differential equations is given. Method of analytical solution of multipoint boundary problems in structural analysis is offered as well. Its major peculiarities include universality, computer-oriented algorithm involving theory of distributions, computational stability, optimal conditionality of resultant systems, partial Jordan decomposition of matrix of coefficients, eliminating necessity of calculation of root vectors. Brief information concerning developed software is provided.
In this paper we consider modelling of composite material with inclusions where the elastic material properties of both matrix and inclusions are uncertain and vary within prescribed bounds. Such mechanical systems, involving interval uncertainties and modelled by finite element method, can be described by parameter dependent systems of linear interval equations and process variables depending on the system solution. A newly developed hybrid interval approach for solving parametric interval linear systems is applied to the considered model and the results are compared to other interval methods. The hybrid approach provides very sharp bounds for the process variables - element strains and stresses. The sources for overestimation when dealing with interval computations are demonstrated. Based on the element strains and stresses, we introduce a definition for the values of nodal strains and stresses by using a set-theoretic approach.
Die Methode der Finiten Elemente ist ein numerisches Verfahren zur Interpolation vorgegebener Werte und zur numerischen Approximation von Lösungen stationärer oder instationärer partieller Differentialgleichungen bzw. Systemen partieller Differentialgleichungen. Grundlage dieser Verfahren ist die Formulierung geeigneter Finiter Elemente und Finiter Element Zerlegungen. Finite Elemente besitzen in der Regel eine geometrische Basis bestehend aus Strecken im eindimensionalen, Drei- oder Vierecken im zweidimensionalen und Tetra- oder Hexaedern im dreidimensionalen euklidischen Raum, eine Menge von Freiheitsgraden und eine Basis von Funktionen. Die geometrische Basis eines Finiten Elements wird verallgemeinert als geometrische Zelle formuliert. Diese geschlossene geometrische Formulierung führt zu einer geometrieunabhängigen Definition der Basisfunktionen eines Finiten Elements in den Zellkoordinaten der geometrischen Zelle. Finite Elemente auf der Basis geometrischer Zellen werden als Bestandteile Finiter Element Zerlegungen in Finiten Element Interpolationen und Finiten Element Approximationen verwendet. Die Finiten Element Approximationen werden am Beispiel der 2-dimensionalen Diffusionsgleichung über das Standard-Galerkin-Verfahren ermittelt.
Hydro- und morphodynamischen Prozesse in Binnengewässern und im Küstennahbereich erzeugen hochkomplexe Phänomene. Zur Beurteilung der Entwicklung von Küstenzohnen, von Flussbetten sowie von Eingriffen des Menschen in Form von Schutzbauwerken sind geeignete numerische Modellwerkzeuge notwendig. Es wird ein holistischer Modellansatz zur Approximation gekoppelter Seegangs-, Strömungs- und Morphodynamischer Prozesse auf der Basis stabilisierter Finiter Elemente vorgestellt. Der Großteil der Modellgleichungen der Hydro- und Morphodynamik sind Transportgleichungen. Dem Transportcharakter dieser Gleichungen entsprechend wird ein stabilisiertes Finites Element Verfahren auf Dreiecken vorgestellt. Die vorgestellte Approximation entspricht einem streamline upwinding Petrov-Galerkin-Verfahrens für vektorwertige mehrdimensionale Probleme, bei dem der Fehler eines Standard-Galerkin-Verfahrens mit Hilfe eines Upwinding-Koeffizienten minimiert wird. Die Wahl des Upwinding-Koeffizienten ist übertragbar auf andere Problemklassen und basiert ausschließlich auf dem Charakter der zugrundeliegene Das Modell wurde für Seegangs- und Strömungs-Untersuchungen im Jade-Weser-Ästuar an der deutschen Nordseeküste eingesetzt.
A geometrical inclusion-matrix model for the finite element analysis of concrete at multiple scales
(2003)
This paper introduces a method to generate adequate inclusion-matrix geometries of concrete in two and three dimensions, which are independent of any specific numerical discretization. The article starts with an analysis on shapes of natural aggregates and discusses corresponding mathematical realizations. As a first prototype a two-dimensional generation of a mesoscale model is introduced. Particle size distribution functions are analysed and prepared for simulating an adequate three-dimensional representation of the aggregates within a concrete structure. A sample geometry of a three-dimensional test cube is generated and the finite element analysis of its heterogeneous geometry by a uniform mesh is presented. Concluding, aspects of a multiscale analysis are discussed and possible enhancements are proposed.
Iso-parametric finite elements with linear shape functions show in general a too stiff element behavior, called locking. By the investigation of structural parts under bending loading the so-called shear locking appears, because these elements can not reproduce pure bending modes. Many studies dealt with the locking problem and a number of methods to avoid the undesirable effects have been developed. Two well known methods are the >Assumed Natural Strain< (ANS) method and the >Enhanced Assumed Strain< (EAS) method. In this study the EAS method is applied to a four-node plane element with four EAS-parameters. The paper will describe the well-known linear formulation, its extension to nonlinear materials and the modeling of material uncertainties with random fields. For nonlinear material behavior the EAS parameters can not be determined directly. Here the problem is solved by using an internal iteration at the element level, which is much more efficient and stable than the determination via a global iteration. To verify the deterministic element behavior the results of common test examples are presented for linear and nonlinear materials. The modeling of material uncertainties is done by point-discretized random fields. To show the applicability of the element for stochastic finite element calculations Latin Hypercube Sampling was applied to investigate the stochastic hardening behavior of a cantilever beam with nonlinear material. The enhanced linear element can be applied as an alternative to higher-order finite elements where more nodes are necessary. The presented element formulation can be used in a similar manner to improve stochastic linear solid elements.
A realistic and reliable model is an important precondition for the simulation of revitalization tasks and the estimation of system properties of existing buildings. Thereby, the main focus lies on the parameter identification, the optimization strategies and the preparation of experiments. As usual structures are modeled by the finite element method. This as well as other techniques are based on idealizations and empiric material properties. Within one theory the parameters of the model should be approximated by gradually performed experiments and their analysis. This approximation method is performed by solving an optimization problem, which is usually non-convex, of high dimension and possesses a non-differentiable objective function. Therefore we use an optimization procedure based on genetic algorithms which was implemented by using the program package SLang...
The worldwide growth of communication networks and associated technologies provide the basic infrastructure for new ways of executing the engineering process. Collaboration amongst team members seperated in time and location is of particular importance. Two broad themes can be recognized in research pertaining to distributed collaboration. One theme focusses on the technical and technological aspects of distributed work, while the other emphasises human aspects thereof. The case of finite element structural analysis in a distributed collaboratory is examined in this paper. An approach is taken which has its roots in human aspects of the structural analysis task. Based on experience of how structural engineers currently approach and execute this task while utilising standard software designed for use on local workstations only, criteria are stated for a software architechture that could support collaborative structural analysis. Aspects of a pilot application and the results of qualitative performance measurements are discussed.
The optimization of continuous structures requires careful attention to discretization errors. Compared to ordinary low order formulation (h-elements) in conjunction with an adaptive mesh refinement in each optimization step, the use of high order finite elements (so called p-elements) has several advantages. However, compared to the h-method a higher order finite element analysis program poses higher demands from a software engineering point of view. In this article the basics of an object oriented higher order finite element system especially tailored to the use in structural optimization is presented. Besides the design of the system, aspects related to the employed implementation language Java are discussed.